diff options
| author | Patrick Steinhardt <psteinhardt@gitlab.com> | 2022-03-18 16:27:33 +0300 |
|---|---|---|
| committer | Patrick Steinhardt <psteinhardt@gitlab.com> | 2022-03-18 16:27:33 +0300 |
| commit | 67a783574950e069587184bcc897bbab1662fe1d (patch) | |
| tree | aaa997f9f1ca208db87132971ac96ee405db5158 /cmd | |
| parent | ff2f4ee8b2ef8f200775d45c550bfb178c55ba9a (diff) | |
housekeeping: Fix Prometheus metrics' latency bucketspks-housekeeping-track-total-count-of-optimizations
The housekeeping manager is hosting a latency metric which tracks how
long specific tasks of the housekeeping manager take. But while latency
metrics should typically use the GRPC latency buckets as configured in
the Gitaly configuration, we accidentally didn't in the context of the
housekeeping manager.
Fix this bug by passing in the Prometheus configuration.
Diffstat (limited to 'cmd')
| -rw-r--r-- | cmd/gitaly/main.go | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/cmd/gitaly/main.go b/cmd/gitaly/main.go index f0b525da2..7d58bdd88 100644 --- a/cmd/gitaly/main.go +++ b/cmd/gitaly/main.go @@ -166,7 +166,7 @@ func run(cfg config.Cfg) error { transactionManager := transaction.NewManager(cfg, registry) prometheus.MustRegister(transactionManager) - housekeepingManager := housekeeping.NewManager(transactionManager) + housekeepingManager := housekeeping.NewManager(cfg.Prometheus, transactionManager) prometheus.MustRegister(housekeepingManager) hookManager := hook.Manager(hook.DisabledManager{}) |