Increase the buckets for compaction duration (#8342)

We're seeing compactions that are taking hours in Cortex which this is
missing. I know while it is not common in Prometheus, I am pretty sure
there are setups where compaction takes longer than 512s. On our own
Prometheus the average compaction duration is 566s.

Signed-off-by: Goutham Veeramachaneni <gouthamve@gmail.com>
This commit is contained in:
Goutham Veeramachaneni 2021-01-25 19:12:18 +01:00 committed by GitHub
parent b5dfa2414b
commit e572e77d90
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23

View file

@ -111,7 +111,7 @@ func newCompactorMetrics(r prometheus.Registerer) *compactorMetrics {
m.duration = prometheus.NewHistogram(prometheus.HistogramOpts{
Name: "prometheus_tsdb_compaction_duration_seconds",
Help: "Duration of compaction runs",
Buckets: prometheus.ExponentialBuckets(1, 2, 10),
Buckets: prometheus.ExponentialBuckets(1, 2, 14),
})
m.chunkSize = prometheus.NewHistogram(prometheus.HistogramOpts{
Name: "prometheus_tsdb_compaction_chunk_size_bytes",