GitLab

Increment multi-zone storage nodes by 1 to 34

See merge request !1917

Sets initial node size to 2 on the default node pool

See merge request !1916

Resources for external-dns in ops

See merge request !1915

Add a service account, intended to be used via GKE workload identity.
Bind dns.admin role to it. Reading
https://cloud.google.com/dns/docs/access-control, it appears as if there
is no way to scope Cloud DNS permissions by zone.

Add a zone in which external-dns will create records, and delegate to it
via an NS record in the gitlab.net root zone in Cloudflare.

Add Thanos Rule to gprd AM allow

See merge request !1922

Add the ops Thanos Rule server subnet to the gprd alertmanager allow
firewall list.
* Fix typo in ops prometheus comment.

https://gitlab.com/gitlab-com/gl-infra/infrastructure/-/issues/10801Signed-off-by: Ben Kochie <[email protected]>

Update dns record for vault-nonprod.gitlab.net

See merge request !1920

This is part of https://gitlab.com/gitlab-com/gl-infra/infrastructure/-/issues/7249

With a small change in network topology for vault-nonprod to allow more complex
VPC peering across environments (for Vault access), vault-nonprod had its ingress
redeployed and thus the IP changed.

Revert "Revert "Merge branch 'infra-10689' into 'master'""

See merge request !1919

add labels to runner managers for accounting purposes

See merge request !1918

This reverts commit 443c4232.

* This is to ensure clean terraform runs
* During incident !2381 this value was changed, and is immutable
* To avoid having to recreate this node pool, this change will allow
terraform runs to pass without requiring a recreation of this node pool

remove central networks as we have decided not to use central region

See merge request !1861

Add vault development server address

See merge request !1914

Allow vault-nonprod to peer with non production environments

See merge request !1906

Prepare lb-registry-temp nodes for production use

See merge request !1913

We created two "temporary" nodes to address
https://gitlab.com/gitlab-com/gl-infra/production/-/issues/2381.
Switching traffic to them restored service. We aren't confident in
re-routing traffic to the nodes in the old module, so we'll keep using
this new one for production traffic. For that purpose, we:
- Distribute nodes amongst availability zones
- Overscale preemptively
- Persist the new state of the registry load balancer

Revert "Merge branch 'infra-10689' into 'master'"

See merge request !1911

This reverts merge request !1868

Add a temporary fe-lb-registry nodes

See merge request !1910

Related to https://gitlab.com/gitlab-com/gl-infra/production/-/issues/2381

remove CI HAProxy nodes from Terraform

See merge request !1868

Reference: https://gitlab.com/gitlab-com/gl-infra/infrastructure/-/issues/10689

Revert "Add a temporary Patroni node"

See merge request !1907

This reverts commit b7810cdf.

Enable slowlog reporting on praefects CloudSQL

See merge request !1905

This is part of https://gitlab.com/gitlab-com/gl-infra/infrastructure/-/issues/7249

This is needed so non-production environments can access the non-production
Vault environment. Ops needs access for CI jobs run against vault-nonprod.

While debugging https://gitlab.com/gitlab-org/gitlab/-/issues/227215 it became apparent that the CloudSQL instance had been running at 100% CPU, and we didn't know what queries were slow.  We set log_min_duration_statement by hand, and then ret-conned this MR to match.  1000ms seems like a reasonable starting point, we can adjust later as required (the number is not particularly special, just one that feels like it will show us really slow queries without swamping us with noise).  Once we get things under control, lowering it is a quite reasonable course of action.

Add/update entries for Vault environments

See merge request !1904

This is part of https://gitlab.com/gitlab-com/gl-infra/infrastructure/-/issues/7249

As we have redeployed production, and deployed non-production for the first time,
the addresses need to be updated/added.

Add a temporary Patroni node

See merge request !1903

Part of debugging https://gitlab.com/gitlab-com/gl-infra/infrastructure/-/issues/10172

resize default node pool in pre

See merge request !1902

Signed-off-by: Michal Wasilewski <[email protected]>

resize default pool

See merge request !1901

Signed-off-by: Michal Wasilewski <[email protected]>

work around enable_flow_logs bug in google provider 2.x

See merge request !1900