Server settings
Datagrok settings come in two kinds:
- Local settings shape the user interface for one user, such as formats, colors, and hidden menus. Every user edits their own. Administrators can set defaults and lock them for groups (see Managed settings).
- Server settings apply to the whole instance, such as authentication, email, logging, and data retention. Only users with the Edit Plugins Settings global permission see and edit them.
To open them, on the Sidebar, click Browse () > Platform > Settings. Local settings are under Local, and server settings are under Server.
Server pages
| Page | What it controls | Learn more |
|---|---|---|
| Admin | Endpoints: the public URL of the instance, used in links and invitations | — |
| E-mail: SMTP, Mailgun, Amazon SES, or the Datagrok mailer | Configure SMTP | |
| Error reporting: where user error reports go, and whether errors are reported automatically | Feedback | |
| AI Providers: the model provider behind AI features | Configure AI providers | |
| Garbage Collector and Maintenance: data retention and the nightly cleanup window | Data retention | |
| Users and Sessions | Sign-up and sign-in rules, LDAP, OpenID, SAML, Google IAP, and group synchronization | Configure authentication |
| Logger | Which events are logged for whom, and Log sync to Amazon CloudWatch, Google Cloud Logging, or OpenTelemetry | Audit, Export logs |
| Scripting | The script languages enabled on the instance | Scripting |
| Cache | Server-side and client-side caching of function results and files, on or off for the whole instance | Cache function results |
| Global Permissions | What each group or role may do across the platform. Requires Edit Global Permissions | Global permissions |
Two more nodes sit next to these pages under Settings:
- Keys: the server keys that encrypt stored credentials and sign sign-in tokens. Requires Admin Keys. See Server keys.
- Repositories: the package registries the server installs plugins from. Requires Browse Plugins.
Data retention
The server deletes old log records, sessions, test runs, and orphaned data in a nightly cleanup, the garbage collector. It is off by default, so on a new instance nothing is deleted and the database keeps growing. Turn it on in Admin > Garbage Collector > Enabled.
| Setting | Default |
|---|---|
| Audit records (the oldest record of each entity is kept) | 365 days |
| Sessions | 365 days |
| Errors, warnings, info, and usage logs | 183 days |
| Unused versions of locally published packages that are not current | 183 days |
| Test runs | 90 days |
| Debug logs, HTTP request statistics, empty sessions, expired sign-in codes | 30 days |
| Orphaned tables and script runs | Seven days |
The cleanup starts on the Maintenance schedule (cron, UTC, default
0 0 * * *, midnight every day) and runs for at most the maintenance duration
(default three hours). The Garbage Collector entry of the health check
reports errors from the last run.
If your policy requires keeping audit records longer than the database does, export them with Log sync and keep them in your log archive.
Set settings in code
Server settings can be part of your infrastructure code instead of the UI.
On any server settings page, click {} to get the settings as JSON, then put
them in the settings section of GROK_PARAMETERS. See
Server configuration.
On instances managed by an external admin console, such as Datagrok SaaS, some server pages are read-only and show Managed by your administrator.
See also: