GitLab

GitLab

Coding agents tell you what they cost and how widely they are used. They cannot tell you whether the work actually ships. That signal lives in GitLab: when a merge request opens, how long it waits for review, how often it comes back, and when it finally merges.

Connecting GitLab gives Dash0 the delivery half of the picture. Pair it with at least one coding agent and Dash0 can compare assisted and non-assisted work side by side, instead of reporting token spend that nothing ties to an outcome.

Use Cases

• Track MR cycle time: Measure how long changes wait between opening, first review, and merge, and see where the time actually goes.
• Monitor pipeline flow: Follow review rounds, rework, and merge frequency to find the stages that hold delivery back.
• Compare assisted and non-assisted work: Correlate merged MRs with coding agent activity to see what difference the agents make.
• Understand team throughput: Analyze delivery per repository and per team to ground engineering decisions in what shipped.

Overview

Dash0 reads delivery activity through group webhooks on GitLab.com or a self-managed instance. There is nothing to deploy: GitLab sends merge request and pipeline events to Dash0, and AI SDLC Insights reports on them.

Prerequisites

• Admin permissions in Dash0.
• GitLab 19.3 or later, using the standard webhook format.
• For self-managed instances, HTTPS on port 443 with no installation path prefix for MR classification.
• A GitLab service account with a read_api access token and access to the groups you want to connect.
• GitLab Premium or Ultimate for group webhooks, on GitLab.com and self-managed instances.
• Optional: a connected coding agent, such as Claude Code or Cursor. GitLab on its own reports delivery metrics; a coding agent adds the comparison between assisted and non-assisted work.

Connect GitLab

1. Open Settings → Integrations, click + Add, and select GitLab.
2. Choose GitLab.com (cloud) or Self-managed. For self-managed, supply the instance URL and OAuth application details from the GitLab organization setup guide.
3. In GitLab, open Manage → Members for each group or project whose repositories you want to include. Invite the service account as a member with the Developer role. Creating a service account grants it no access, and a private group is invisible to a non-member.
4. In Dash0, add the groups containing the repositories you want to include, with a service-account token for each group. Dash0 can access only repositories that the service account can access through its group or project memberships. Invite it to each private group you add so Dash0 can see the group.

Select datasets

Select the datasets in which the MRs should appear, then click Connect. If you use coding agents, pick the dataset they send to: Dash0 compares assisted and non-assisted work only within one dataset. Without coding agents, any dataset works. You can edit the datasets later in Settings → Integrations.

Configure webhooks

For each connected group, follow Add the webhook to add the integration's webhook URL in GitLab, then select Generate signing token in GitLab and paste the token into Dash0.

Under Trigger, select these events:

• Push events, with All branches
• Comments
• Merge request events
• Job events
• Pipeline events
• Emoji events

Dash0 ignores every other trigger, so you can leave the rest cleared. Don't use a custom webhook template: Dash0 needs GitLab's standard payload format.

To recover missed events, use GitLab's webhook delivery history to resend recent deliveries. For backfill from events Dash0 already received, contact Dash0 support to replay retained events.

Verify

Open and merge an MR in a connected repository, then open AI SDLC Insights in Dash0 in the selected dataset. Check that the MR contributes to delivery metrics. Cycle time and merge counts need a merged MR, so an open one alone leaves them empty.