databricks-cli/cmd/auth/login.go

224 lines
7.7 KiB
Go
Raw Permalink Normal View History

package auth
import (
"context"
"errors"
"fmt"
"runtime"
"time"
"github.com/databricks/cli/libs/auth"
"github.com/databricks/cli/libs/cmdio"
"github.com/databricks/cli/libs/databrickscfg"
"github.com/databricks/cli/libs/databrickscfg/cfgpickers"
Improve token refresh flow (#1434) ## Changes Currently, there are a number of issues with the non-happy-path flows for token refresh in the CLI. If the token refresh fails, the raw error message is presented to the user, as seen below. This message is very difficult for users to interpret and doesn't give any clear direction on how to resolve this issue. ``` Error: token refresh: Post "https://adb-<WSID>.azuredatabricks.net/oidc/v1/token": http 400: {"error":"invalid_request","error_description":"Refresh token is invalid"} ``` When logging in again, I've noticed that the timeout for logging in is very short, only 45 seconds. If a user is using a password manager and needs to login to that first, or needs to do MFA, 45 seconds may not be enough time. to an account-level profile, it is quite frustrating for users to need to re-enter account ID information when that information is already stored in the user's `.databrickscfg` file. This PR tackles these two issues. First, the presentation of error messages from `databricks auth token` is improved substantially by converting the `error` into a human-readable message. When the refresh token is invalid, it will present a command for the user to run to reauthenticate. If the token fetching failed for some other reason, that reason will be presented in a nice way, providing front-line debugging steps and ultimately redirecting users to file a ticket at this repo if they can't resolve the issue themselves. After this PR, the new error message is: ``` Error: a new access token could not be retrieved because the refresh token is invalid. To reauthenticate, run `.databricks/databricks auth login --host https://adb-<WSID>.azuredatabricks.net` ``` To improve the login flow, this PR modifies `databricks auth login` to auto-complete the account ID from the profile when present. Additionally, it increases the login timeout from 45 seconds to 1 hour to give the user sufficient time to login as needed. To test this change, I needed to refactor some components of the CLI around profile management, the token cache, and the API client used to fetch OAuth tokens. These are now settable in the context, and a demonstration of how they can be set and used is found in `auth_test.go`. Separately, this also demonstrates a sort-of integration test of the CLI by executing the Cobra command for `databricks auth token` from tests, which may be useful for testing other end-to-end functionality in the CLI. In particular, I believe this is necessary in order to set flag values (like the `--profile` flag in this case) for use in testing. ## Tests Unit tests cover the unhappy and happy paths using the mocked API client, token cache, and profiler. Manually tested --------- Co-authored-by: Pieter Noordhuis <pieter.noordhuis@databricks.com>
2024-05-16 10:22:09 +00:00
"github.com/databricks/cli/libs/databrickscfg/profile"
"github.com/databricks/databricks-sdk-go"
"github.com/databricks/databricks-sdk-go/config"
"github.com/spf13/cobra"
)
func promptForProfile(ctx context.Context, defaultValue string) (string, error) {
if !cmdio.IsInTTY(ctx) {
return "", nil
}
prompt := cmdio.Prompt(ctx)
prompt.Label = "Databricks profile name"
prompt.Default = defaultValue
prompt.AllowEdit = true
return prompt.Run()
}
const minimalDbConnectVersion = "13.1"
Improve token refresh flow (#1434) ## Changes Currently, there are a number of issues with the non-happy-path flows for token refresh in the CLI. If the token refresh fails, the raw error message is presented to the user, as seen below. This message is very difficult for users to interpret and doesn't give any clear direction on how to resolve this issue. ``` Error: token refresh: Post "https://adb-<WSID>.azuredatabricks.net/oidc/v1/token": http 400: {"error":"invalid_request","error_description":"Refresh token is invalid"} ``` When logging in again, I've noticed that the timeout for logging in is very short, only 45 seconds. If a user is using a password manager and needs to login to that first, or needs to do MFA, 45 seconds may not be enough time. to an account-level profile, it is quite frustrating for users to need to re-enter account ID information when that information is already stored in the user's `.databrickscfg` file. This PR tackles these two issues. First, the presentation of error messages from `databricks auth token` is improved substantially by converting the `error` into a human-readable message. When the refresh token is invalid, it will present a command for the user to run to reauthenticate. If the token fetching failed for some other reason, that reason will be presented in a nice way, providing front-line debugging steps and ultimately redirecting users to file a ticket at this repo if they can't resolve the issue themselves. After this PR, the new error message is: ``` Error: a new access token could not be retrieved because the refresh token is invalid. To reauthenticate, run `.databricks/databricks auth login --host https://adb-<WSID>.azuredatabricks.net` ``` To improve the login flow, this PR modifies `databricks auth login` to auto-complete the account ID from the profile when present. Additionally, it increases the login timeout from 45 seconds to 1 hour to give the user sufficient time to login as needed. To test this change, I needed to refactor some components of the CLI around profile management, the token cache, and the API client used to fetch OAuth tokens. These are now settable in the context, and a demonstration of how they can be set and used is found in `auth_test.go`. Separately, this also demonstrates a sort-of integration test of the CLI by executing the Cobra command for `databricks auth token` from tests, which may be useful for testing other end-to-end functionality in the CLI. In particular, I believe this is necessary in order to set flag values (like the `--profile` flag in this case) for use in testing. ## Tests Unit tests cover the unhappy and happy paths using the mocked API client, token cache, and profiler. Manually tested --------- Co-authored-by: Pieter Noordhuis <pieter.noordhuis@databricks.com>
2024-05-16 10:22:09 +00:00
const defaultTimeout = 1 * time.Hour
func newLoginCommand(persistentAuth *auth.PersistentAuth) *cobra.Command {
defaultConfigPath := "~/.databrickscfg"
if runtime.GOOS == "windows" {
defaultConfigPath = "%USERPROFILE%\\.databrickscfg"
}
cmd := &cobra.Command{
Use: "login [HOST]",
Short: "Log into a Databricks workspace or account",
Long: fmt.Sprintf(`Log into a Databricks workspace or account.
This command logs you into the Databricks workspace or account and saves
the authentication configuration in a profile (in %s by default).
This profile can then be used to authenticate other Databricks CLI commands by
specifying the --profile flag. This profile can also be used to authenticate
other Databricks tooling that supports the Databricks Unified Authentication
Specification. This includes the Databricks Go, Python, and Java SDKs. For more information,
you can refer to the documentation linked below.
AWS: https://docs.databricks.com/dev-tools/auth/index.html
Azure: https://learn.microsoft.com/azure/databricks/dev-tools/auth
GCP: https://docs.gcp.databricks.com/dev-tools/auth/index.html
This command requires a Databricks Host URL (using --host or as a positional argument
or implicitly inferred from the specified profile name)
and a profile name (using --profile) to be specified. If you don't specify these
values, you'll be prompted for values at runtime.
While this command always logs you into the specified host, the runtime behaviour
depends on the existing profiles you have set in your configuration file
(at %s by default).
1. If a profile with the specified name exists and specifies a host, you'll
be logged into the host specified by the profile. The profile will be updated
to use "databricks-cli" as the auth type if that was not the case before.
2. If a profile with the specified name exists but does not specify a host,
you'll be prompted to specify a host. The profile will be updated to use the
specified host. The auth type will be updated to "databricks-cli" if that was
not the case before.
3. If a profile with the specified name exists and specifies a host, but you
specify a host using --host (or as the [HOST] positional arg), the profile will
be updated to use the newly specified host. The auth type will be updated to
"databricks-cli" if that was not the case before.
4. If a profile with the specified name does not exist, a new profile will be
created with the specified host. The auth type will be set to "databricks-cli".
`, defaultConfigPath, defaultConfigPath),
}
var loginTimeout time.Duration
var configureCluster bool
Improve token refresh flow (#1434) ## Changes Currently, there are a number of issues with the non-happy-path flows for token refresh in the CLI. If the token refresh fails, the raw error message is presented to the user, as seen below. This message is very difficult for users to interpret and doesn't give any clear direction on how to resolve this issue. ``` Error: token refresh: Post "https://adb-<WSID>.azuredatabricks.net/oidc/v1/token": http 400: {"error":"invalid_request","error_description":"Refresh token is invalid"} ``` When logging in again, I've noticed that the timeout for logging in is very short, only 45 seconds. If a user is using a password manager and needs to login to that first, or needs to do MFA, 45 seconds may not be enough time. to an account-level profile, it is quite frustrating for users to need to re-enter account ID information when that information is already stored in the user's `.databrickscfg` file. This PR tackles these two issues. First, the presentation of error messages from `databricks auth token` is improved substantially by converting the `error` into a human-readable message. When the refresh token is invalid, it will present a command for the user to run to reauthenticate. If the token fetching failed for some other reason, that reason will be presented in a nice way, providing front-line debugging steps and ultimately redirecting users to file a ticket at this repo if they can't resolve the issue themselves. After this PR, the new error message is: ``` Error: a new access token could not be retrieved because the refresh token is invalid. To reauthenticate, run `.databricks/databricks auth login --host https://adb-<WSID>.azuredatabricks.net` ``` To improve the login flow, this PR modifies `databricks auth login` to auto-complete the account ID from the profile when present. Additionally, it increases the login timeout from 45 seconds to 1 hour to give the user sufficient time to login as needed. To test this change, I needed to refactor some components of the CLI around profile management, the token cache, and the API client used to fetch OAuth tokens. These are now settable in the context, and a demonstration of how they can be set and used is found in `auth_test.go`. Separately, this also demonstrates a sort-of integration test of the CLI by executing the Cobra command for `databricks auth token` from tests, which may be useful for testing other end-to-end functionality in the CLI. In particular, I believe this is necessary in order to set flag values (like the `--profile` flag in this case) for use in testing. ## Tests Unit tests cover the unhappy and happy paths using the mocked API client, token cache, and profiler. Manually tested --------- Co-authored-by: Pieter Noordhuis <pieter.noordhuis@databricks.com>
2024-05-16 10:22:09 +00:00
cmd.Flags().DurationVar(&loginTimeout, "timeout", defaultTimeout,
"Timeout for completing login challenge in the browser")
cmd.Flags().BoolVar(&configureCluster, "configure-cluster", false,
"Prompts to configure cluster")
cmd.RunE = func(cmd *cobra.Command, args []string) error {
ctx := cmd.Context()
profileName := cmd.Flag("profile").Value.String()
// If the user has not specified a profile name, prompt for one.
if profileName == "" {
var err error
profileName, err = promptForProfile(ctx, persistentAuth.ProfileName())
if err != nil {
return err
}
}
// Set the host and account-id based on the provided arguments and flags.
Improve token refresh flow (#1434) ## Changes Currently, there are a number of issues with the non-happy-path flows for token refresh in the CLI. If the token refresh fails, the raw error message is presented to the user, as seen below. This message is very difficult for users to interpret and doesn't give any clear direction on how to resolve this issue. ``` Error: token refresh: Post "https://adb-<WSID>.azuredatabricks.net/oidc/v1/token": http 400: {"error":"invalid_request","error_description":"Refresh token is invalid"} ``` When logging in again, I've noticed that the timeout for logging in is very short, only 45 seconds. If a user is using a password manager and needs to login to that first, or needs to do MFA, 45 seconds may not be enough time. to an account-level profile, it is quite frustrating for users to need to re-enter account ID information when that information is already stored in the user's `.databrickscfg` file. This PR tackles these two issues. First, the presentation of error messages from `databricks auth token` is improved substantially by converting the `error` into a human-readable message. When the refresh token is invalid, it will present a command for the user to run to reauthenticate. If the token fetching failed for some other reason, that reason will be presented in a nice way, providing front-line debugging steps and ultimately redirecting users to file a ticket at this repo if they can't resolve the issue themselves. After this PR, the new error message is: ``` Error: a new access token could not be retrieved because the refresh token is invalid. To reauthenticate, run `.databricks/databricks auth login --host https://adb-<WSID>.azuredatabricks.net` ``` To improve the login flow, this PR modifies `databricks auth login` to auto-complete the account ID from the profile when present. Additionally, it increases the login timeout from 45 seconds to 1 hour to give the user sufficient time to login as needed. To test this change, I needed to refactor some components of the CLI around profile management, the token cache, and the API client used to fetch OAuth tokens. These are now settable in the context, and a demonstration of how they can be set and used is found in `auth_test.go`. Separately, this also demonstrates a sort-of integration test of the CLI by executing the Cobra command for `databricks auth token` from tests, which may be useful for testing other end-to-end functionality in the CLI. In particular, I believe this is necessary in order to set flag values (like the `--profile` flag in this case) for use in testing. ## Tests Unit tests cover the unhappy and happy paths using the mocked API client, token cache, and profiler. Manually tested --------- Co-authored-by: Pieter Noordhuis <pieter.noordhuis@databricks.com>
2024-05-16 10:22:09 +00:00
err := setHostAndAccountId(ctx, profileName, persistentAuth, args)
if err != nil {
return err
}
defer persistentAuth.Close()
// We need the config without the profile before it's used to initialise new workspace client below.
// Otherwise it will complain about non existing profile because it was not yet saved.
cfg := config.Config{
Improve token refresh flow (#1434) ## Changes Currently, there are a number of issues with the non-happy-path flows for token refresh in the CLI. If the token refresh fails, the raw error message is presented to the user, as seen below. This message is very difficult for users to interpret and doesn't give any clear direction on how to resolve this issue. ``` Error: token refresh: Post "https://adb-<WSID>.azuredatabricks.net/oidc/v1/token": http 400: {"error":"invalid_request","error_description":"Refresh token is invalid"} ``` When logging in again, I've noticed that the timeout for logging in is very short, only 45 seconds. If a user is using a password manager and needs to login to that first, or needs to do MFA, 45 seconds may not be enough time. to an account-level profile, it is quite frustrating for users to need to re-enter account ID information when that information is already stored in the user's `.databrickscfg` file. This PR tackles these two issues. First, the presentation of error messages from `databricks auth token` is improved substantially by converting the `error` into a human-readable message. When the refresh token is invalid, it will present a command for the user to run to reauthenticate. If the token fetching failed for some other reason, that reason will be presented in a nice way, providing front-line debugging steps and ultimately redirecting users to file a ticket at this repo if they can't resolve the issue themselves. After this PR, the new error message is: ``` Error: a new access token could not be retrieved because the refresh token is invalid. To reauthenticate, run `.databricks/databricks auth login --host https://adb-<WSID>.azuredatabricks.net` ``` To improve the login flow, this PR modifies `databricks auth login` to auto-complete the account ID from the profile when present. Additionally, it increases the login timeout from 45 seconds to 1 hour to give the user sufficient time to login as needed. To test this change, I needed to refactor some components of the CLI around profile management, the token cache, and the API client used to fetch OAuth tokens. These are now settable in the context, and a demonstration of how they can be set and used is found in `auth_test.go`. Separately, this also demonstrates a sort-of integration test of the CLI by executing the Cobra command for `databricks auth token` from tests, which may be useful for testing other end-to-end functionality in the CLI. In particular, I believe this is necessary in order to set flag values (like the `--profile` flag in this case) for use in testing. ## Tests Unit tests cover the unhappy and happy paths using the mocked API client, token cache, and profiler. Manually tested --------- Co-authored-by: Pieter Noordhuis <pieter.noordhuis@databricks.com>
2024-05-16 10:22:09 +00:00
Host: persistentAuth.Host,
AccountID: persistentAuth.AccountID,
AuthType: "databricks-cli",
}
ctx, cancel := context.WithTimeout(ctx, loginTimeout)
defer cancel()
err = persistentAuth.Challenge(ctx)
if err != nil {
return err
}
if configureCluster {
w, err := databricks.NewWorkspaceClient((*databricks.Config)(&cfg))
if err != nil {
return err
}
ctx := cmd.Context()
clusterID, err := cfgpickers.AskForCluster(ctx, w,
cfgpickers.WithDatabricksConnect(minimalDbConnectVersion))
if err != nil {
return err
}
cfg.ClusterID = clusterID
}
if profileName != "" {
err = databrickscfg.SaveToProfile(ctx, &config.Config{
Profile: profileName,
Host: cfg.Host,
AuthType: cfg.AuthType,
AccountID: cfg.AccountID,
ClusterID: cfg.ClusterID,
})
if err != nil {
return err
}
cmdio.LogString(ctx, fmt.Sprintf("Profile %s was successfully saved", profileName))
}
return nil
}
return cmd
}
// Sets the host in the persistentAuth object based on the provided arguments and flags.
// Follows the following precedence:
// 1. [HOST] (first positional argument) or --host flag. Error if both are specified.
// 2. Profile host, if available.
// 3. Prompt the user for the host.
//
// Set the account in the persistentAuth object based on the flags.
// Follows the following precedence:
// 1. --account-id flag.
// 2. account-id from the specified profile, if available.
// 3. Prompt the user for the account-id.
Improve token refresh flow (#1434) ## Changes Currently, there are a number of issues with the non-happy-path flows for token refresh in the CLI. If the token refresh fails, the raw error message is presented to the user, as seen below. This message is very difficult for users to interpret and doesn't give any clear direction on how to resolve this issue. ``` Error: token refresh: Post "https://adb-<WSID>.azuredatabricks.net/oidc/v1/token": http 400: {"error":"invalid_request","error_description":"Refresh token is invalid"} ``` When logging in again, I've noticed that the timeout for logging in is very short, only 45 seconds. If a user is using a password manager and needs to login to that first, or needs to do MFA, 45 seconds may not be enough time. to an account-level profile, it is quite frustrating for users to need to re-enter account ID information when that information is already stored in the user's `.databrickscfg` file. This PR tackles these two issues. First, the presentation of error messages from `databricks auth token` is improved substantially by converting the `error` into a human-readable message. When the refresh token is invalid, it will present a command for the user to run to reauthenticate. If the token fetching failed for some other reason, that reason will be presented in a nice way, providing front-line debugging steps and ultimately redirecting users to file a ticket at this repo if they can't resolve the issue themselves. After this PR, the new error message is: ``` Error: a new access token could not be retrieved because the refresh token is invalid. To reauthenticate, run `.databricks/databricks auth login --host https://adb-<WSID>.azuredatabricks.net` ``` To improve the login flow, this PR modifies `databricks auth login` to auto-complete the account ID from the profile when present. Additionally, it increases the login timeout from 45 seconds to 1 hour to give the user sufficient time to login as needed. To test this change, I needed to refactor some components of the CLI around profile management, the token cache, and the API client used to fetch OAuth tokens. These are now settable in the context, and a demonstration of how they can be set and used is found in `auth_test.go`. Separately, this also demonstrates a sort-of integration test of the CLI by executing the Cobra command for `databricks auth token` from tests, which may be useful for testing other end-to-end functionality in the CLI. In particular, I believe this is necessary in order to set flag values (like the `--profile` flag in this case) for use in testing. ## Tests Unit tests cover the unhappy and happy paths using the mocked API client, token cache, and profiler. Manually tested --------- Co-authored-by: Pieter Noordhuis <pieter.noordhuis@databricks.com>
2024-05-16 10:22:09 +00:00
func setHostAndAccountId(ctx context.Context, profileName string, persistentAuth *auth.PersistentAuth, args []string) error {
// If both [HOST] and --host are provided, return an error.
if len(args) > 0 && persistentAuth.Host != "" {
return fmt.Errorf("please only provide a host as an argument or a flag, not both")
}
Improve token refresh flow (#1434) ## Changes Currently, there are a number of issues with the non-happy-path flows for token refresh in the CLI. If the token refresh fails, the raw error message is presented to the user, as seen below. This message is very difficult for users to interpret and doesn't give any clear direction on how to resolve this issue. ``` Error: token refresh: Post "https://adb-<WSID>.azuredatabricks.net/oidc/v1/token": http 400: {"error":"invalid_request","error_description":"Refresh token is invalid"} ``` When logging in again, I've noticed that the timeout for logging in is very short, only 45 seconds. If a user is using a password manager and needs to login to that first, or needs to do MFA, 45 seconds may not be enough time. to an account-level profile, it is quite frustrating for users to need to re-enter account ID information when that information is already stored in the user's `.databrickscfg` file. This PR tackles these two issues. First, the presentation of error messages from `databricks auth token` is improved substantially by converting the `error` into a human-readable message. When the refresh token is invalid, it will present a command for the user to run to reauthenticate. If the token fetching failed for some other reason, that reason will be presented in a nice way, providing front-line debugging steps and ultimately redirecting users to file a ticket at this repo if they can't resolve the issue themselves. After this PR, the new error message is: ``` Error: a new access token could not be retrieved because the refresh token is invalid. To reauthenticate, run `.databricks/databricks auth login --host https://adb-<WSID>.azuredatabricks.net` ``` To improve the login flow, this PR modifies `databricks auth login` to auto-complete the account ID from the profile when present. Additionally, it increases the login timeout from 45 seconds to 1 hour to give the user sufficient time to login as needed. To test this change, I needed to refactor some components of the CLI around profile management, the token cache, and the API client used to fetch OAuth tokens. These are now settable in the context, and a demonstration of how they can be set and used is found in `auth_test.go`. Separately, this also demonstrates a sort-of integration test of the CLI by executing the Cobra command for `databricks auth token` from tests, which may be useful for testing other end-to-end functionality in the CLI. In particular, I believe this is necessary in order to set flag values (like the `--profile` flag in this case) for use in testing. ## Tests Unit tests cover the unhappy and happy paths using the mocked API client, token cache, and profiler. Manually tested --------- Co-authored-by: Pieter Noordhuis <pieter.noordhuis@databricks.com>
2024-05-16 10:22:09 +00:00
profiler := profile.GetProfiler(ctx)
// If the chosen profile has a hostname and the user hasn't specified a host, infer the host from the profile.
Improve token refresh flow (#1434) ## Changes Currently, there are a number of issues with the non-happy-path flows for token refresh in the CLI. If the token refresh fails, the raw error message is presented to the user, as seen below. This message is very difficult for users to interpret and doesn't give any clear direction on how to resolve this issue. ``` Error: token refresh: Post "https://adb-<WSID>.azuredatabricks.net/oidc/v1/token": http 400: {"error":"invalid_request","error_description":"Refresh token is invalid"} ``` When logging in again, I've noticed that the timeout for logging in is very short, only 45 seconds. If a user is using a password manager and needs to login to that first, or needs to do MFA, 45 seconds may not be enough time. to an account-level profile, it is quite frustrating for users to need to re-enter account ID information when that information is already stored in the user's `.databrickscfg` file. This PR tackles these two issues. First, the presentation of error messages from `databricks auth token` is improved substantially by converting the `error` into a human-readable message. When the refresh token is invalid, it will present a command for the user to run to reauthenticate. If the token fetching failed for some other reason, that reason will be presented in a nice way, providing front-line debugging steps and ultimately redirecting users to file a ticket at this repo if they can't resolve the issue themselves. After this PR, the new error message is: ``` Error: a new access token could not be retrieved because the refresh token is invalid. To reauthenticate, run `.databricks/databricks auth login --host https://adb-<WSID>.azuredatabricks.net` ``` To improve the login flow, this PR modifies `databricks auth login` to auto-complete the account ID from the profile when present. Additionally, it increases the login timeout from 45 seconds to 1 hour to give the user sufficient time to login as needed. To test this change, I needed to refactor some components of the CLI around profile management, the token cache, and the API client used to fetch OAuth tokens. These are now settable in the context, and a demonstration of how they can be set and used is found in `auth_test.go`. Separately, this also demonstrates a sort-of integration test of the CLI by executing the Cobra command for `databricks auth token` from tests, which may be useful for testing other end-to-end functionality in the CLI. In particular, I believe this is necessary in order to set flag values (like the `--profile` flag in this case) for use in testing. ## Tests Unit tests cover the unhappy and happy paths using the mocked API client, token cache, and profiler. Manually tested --------- Co-authored-by: Pieter Noordhuis <pieter.noordhuis@databricks.com>
2024-05-16 10:22:09 +00:00
profiles, err := profiler.LoadProfiles(ctx, profile.WithName(profileName))
// Tolerate ErrNoConfiguration here, as we will write out a configuration as part of the login flow.
Improve token refresh flow (#1434) ## Changes Currently, there are a number of issues with the non-happy-path flows for token refresh in the CLI. If the token refresh fails, the raw error message is presented to the user, as seen below. This message is very difficult for users to interpret and doesn't give any clear direction on how to resolve this issue. ``` Error: token refresh: Post "https://adb-<WSID>.azuredatabricks.net/oidc/v1/token": http 400: {"error":"invalid_request","error_description":"Refresh token is invalid"} ``` When logging in again, I've noticed that the timeout for logging in is very short, only 45 seconds. If a user is using a password manager and needs to login to that first, or needs to do MFA, 45 seconds may not be enough time. to an account-level profile, it is quite frustrating for users to need to re-enter account ID information when that information is already stored in the user's `.databrickscfg` file. This PR tackles these two issues. First, the presentation of error messages from `databricks auth token` is improved substantially by converting the `error` into a human-readable message. When the refresh token is invalid, it will present a command for the user to run to reauthenticate. If the token fetching failed for some other reason, that reason will be presented in a nice way, providing front-line debugging steps and ultimately redirecting users to file a ticket at this repo if they can't resolve the issue themselves. After this PR, the new error message is: ``` Error: a new access token could not be retrieved because the refresh token is invalid. To reauthenticate, run `.databricks/databricks auth login --host https://adb-<WSID>.azuredatabricks.net` ``` To improve the login flow, this PR modifies `databricks auth login` to auto-complete the account ID from the profile when present. Additionally, it increases the login timeout from 45 seconds to 1 hour to give the user sufficient time to login as needed. To test this change, I needed to refactor some components of the CLI around profile management, the token cache, and the API client used to fetch OAuth tokens. These are now settable in the context, and a demonstration of how they can be set and used is found in `auth_test.go`. Separately, this also demonstrates a sort-of integration test of the CLI by executing the Cobra command for `databricks auth token` from tests, which may be useful for testing other end-to-end functionality in the CLI. In particular, I believe this is necessary in order to set flag values (like the `--profile` flag in this case) for use in testing. ## Tests Unit tests cover the unhappy and happy paths using the mocked API client, token cache, and profiler. Manually tested --------- Co-authored-by: Pieter Noordhuis <pieter.noordhuis@databricks.com>
2024-05-16 10:22:09 +00:00
if err != nil && !errors.Is(err, profile.ErrNoConfiguration) {
return err
}
Improve token refresh flow (#1434) ## Changes Currently, there are a number of issues with the non-happy-path flows for token refresh in the CLI. If the token refresh fails, the raw error message is presented to the user, as seen below. This message is very difficult for users to interpret and doesn't give any clear direction on how to resolve this issue. ``` Error: token refresh: Post "https://adb-<WSID>.azuredatabricks.net/oidc/v1/token": http 400: {"error":"invalid_request","error_description":"Refresh token is invalid"} ``` When logging in again, I've noticed that the timeout for logging in is very short, only 45 seconds. If a user is using a password manager and needs to login to that first, or needs to do MFA, 45 seconds may not be enough time. to an account-level profile, it is quite frustrating for users to need to re-enter account ID information when that information is already stored in the user's `.databrickscfg` file. This PR tackles these two issues. First, the presentation of error messages from `databricks auth token` is improved substantially by converting the `error` into a human-readable message. When the refresh token is invalid, it will present a command for the user to run to reauthenticate. If the token fetching failed for some other reason, that reason will be presented in a nice way, providing front-line debugging steps and ultimately redirecting users to file a ticket at this repo if they can't resolve the issue themselves. After this PR, the new error message is: ``` Error: a new access token could not be retrieved because the refresh token is invalid. To reauthenticate, run `.databricks/databricks auth login --host https://adb-<WSID>.azuredatabricks.net` ``` To improve the login flow, this PR modifies `databricks auth login` to auto-complete the account ID from the profile when present. Additionally, it increases the login timeout from 45 seconds to 1 hour to give the user sufficient time to login as needed. To test this change, I needed to refactor some components of the CLI around profile management, the token cache, and the API client used to fetch OAuth tokens. These are now settable in the context, and a demonstration of how they can be set and used is found in `auth_test.go`. Separately, this also demonstrates a sort-of integration test of the CLI by executing the Cobra command for `databricks auth token` from tests, which may be useful for testing other end-to-end functionality in the CLI. In particular, I believe this is necessary in order to set flag values (like the `--profile` flag in this case) for use in testing. ## Tests Unit tests cover the unhappy and happy paths using the mocked API client, token cache, and profiler. Manually tested --------- Co-authored-by: Pieter Noordhuis <pieter.noordhuis@databricks.com>
2024-05-16 10:22:09 +00:00
if persistentAuth.Host == "" {
if len(args) > 0 {
// If [HOST] is provided, set the host to the provided positional argument.
persistentAuth.Host = args[0]
} else if len(profiles) > 0 && profiles[0].Host != "" {
// If neither [HOST] nor --host are provided, and the profile has a host, use it.
persistentAuth.Host = profiles[0].Host
} else {
// If neither [HOST] nor --host are provided, and the profile does not have a host,
// then prompt the user for a host.
hostName, err := promptForHost(ctx)
if err != nil {
return err
}
persistentAuth.Host = hostName
}
}
// If the account-id was not provided as a cmd line flag, try to read it from
// the specified profile.
Improve token refresh flow (#1434) ## Changes Currently, there are a number of issues with the non-happy-path flows for token refresh in the CLI. If the token refresh fails, the raw error message is presented to the user, as seen below. This message is very difficult for users to interpret and doesn't give any clear direction on how to resolve this issue. ``` Error: token refresh: Post "https://adb-<WSID>.azuredatabricks.net/oidc/v1/token": http 400: {"error":"invalid_request","error_description":"Refresh token is invalid"} ``` When logging in again, I've noticed that the timeout for logging in is very short, only 45 seconds. If a user is using a password manager and needs to login to that first, or needs to do MFA, 45 seconds may not be enough time. to an account-level profile, it is quite frustrating for users to need to re-enter account ID information when that information is already stored in the user's `.databrickscfg` file. This PR tackles these two issues. First, the presentation of error messages from `databricks auth token` is improved substantially by converting the `error` into a human-readable message. When the refresh token is invalid, it will present a command for the user to run to reauthenticate. If the token fetching failed for some other reason, that reason will be presented in a nice way, providing front-line debugging steps and ultimately redirecting users to file a ticket at this repo if they can't resolve the issue themselves. After this PR, the new error message is: ``` Error: a new access token could not be retrieved because the refresh token is invalid. To reauthenticate, run `.databricks/databricks auth login --host https://adb-<WSID>.azuredatabricks.net` ``` To improve the login flow, this PR modifies `databricks auth login` to auto-complete the account ID from the profile when present. Additionally, it increases the login timeout from 45 seconds to 1 hour to give the user sufficient time to login as needed. To test this change, I needed to refactor some components of the CLI around profile management, the token cache, and the API client used to fetch OAuth tokens. These are now settable in the context, and a demonstration of how they can be set and used is found in `auth_test.go`. Separately, this also demonstrates a sort-of integration test of the CLI by executing the Cobra command for `databricks auth token` from tests, which may be useful for testing other end-to-end functionality in the CLI. In particular, I believe this is necessary in order to set flag values (like the `--profile` flag in this case) for use in testing. ## Tests Unit tests cover the unhappy and happy paths using the mocked API client, token cache, and profiler. Manually tested --------- Co-authored-by: Pieter Noordhuis <pieter.noordhuis@databricks.com>
2024-05-16 10:22:09 +00:00
isAccountClient := (&config.Config{Host: persistentAuth.Host}).IsAccountClient()
if isAccountClient && persistentAuth.AccountID == "" {
if len(profiles) > 0 && profiles[0].AccountID != "" {
persistentAuth.AccountID = profiles[0].AccountID
} else {
// Prompt user for the account-id if it we could not get it from a
// profile.
Improve token refresh flow (#1434) ## Changes Currently, there are a number of issues with the non-happy-path flows for token refresh in the CLI. If the token refresh fails, the raw error message is presented to the user, as seen below. This message is very difficult for users to interpret and doesn't give any clear direction on how to resolve this issue. ``` Error: token refresh: Post "https://adb-<WSID>.azuredatabricks.net/oidc/v1/token": http 400: {"error":"invalid_request","error_description":"Refresh token is invalid"} ``` When logging in again, I've noticed that the timeout for logging in is very short, only 45 seconds. If a user is using a password manager and needs to login to that first, or needs to do MFA, 45 seconds may not be enough time. to an account-level profile, it is quite frustrating for users to need to re-enter account ID information when that information is already stored in the user's `.databrickscfg` file. This PR tackles these two issues. First, the presentation of error messages from `databricks auth token` is improved substantially by converting the `error` into a human-readable message. When the refresh token is invalid, it will present a command for the user to run to reauthenticate. If the token fetching failed for some other reason, that reason will be presented in a nice way, providing front-line debugging steps and ultimately redirecting users to file a ticket at this repo if they can't resolve the issue themselves. After this PR, the new error message is: ``` Error: a new access token could not be retrieved because the refresh token is invalid. To reauthenticate, run `.databricks/databricks auth login --host https://adb-<WSID>.azuredatabricks.net` ``` To improve the login flow, this PR modifies `databricks auth login` to auto-complete the account ID from the profile when present. Additionally, it increases the login timeout from 45 seconds to 1 hour to give the user sufficient time to login as needed. To test this change, I needed to refactor some components of the CLI around profile management, the token cache, and the API client used to fetch OAuth tokens. These are now settable in the context, and a demonstration of how they can be set and used is found in `auth_test.go`. Separately, this also demonstrates a sort-of integration test of the CLI by executing the Cobra command for `databricks auth token` from tests, which may be useful for testing other end-to-end functionality in the CLI. In particular, I believe this is necessary in order to set flag values (like the `--profile` flag in this case) for use in testing. ## Tests Unit tests cover the unhappy and happy paths using the mocked API client, token cache, and profiler. Manually tested --------- Co-authored-by: Pieter Noordhuis <pieter.noordhuis@databricks.com>
2024-05-16 10:22:09 +00:00
accountId, err := promptForAccountID(ctx)
if err != nil {
return err
}
persistentAuth.AccountID = accountId
}
}
return nil
}