Add metric for number of offloaded timelines (#11976)

We want to keep track of the number of offloaded timelines. It's a
per-tenant shard metric because each shard makes offloading decisions on
its own.
This commit is contained in:
Arpad Müller
2025-05-21 13:28:22 +02:00
committed by GitHub
parent 08bb72e516
commit 136cf1979b
4 changed files with 41 additions and 3 deletions

View File

@@ -184,6 +184,7 @@ PAGESERVER_PER_TENANT_METRICS: tuple[str, ...] = (
"pageserver_evictions_with_low_residence_duration_total",
"pageserver_aux_file_estimated_size",
"pageserver_valid_lsn_lease_count",
"pageserver_tenant_offloaded_timelines",
counter("pageserver_tenant_throttling_count_accounted_start"),
counter("pageserver_tenant_throttling_count_accounted_finish"),
counter("pageserver_tenant_throttling_wait_usecs_sum"),

View File

@@ -193,6 +193,11 @@ def test_timeline_offloading(neon_env_builder: NeonEnvBuilder, manual_offload: b
"test_ancestor_branch_archive_branch1", tenant_id, "test_ancestor_branch_archive_parent"
)
offloaded_count = ps_http.get_metric_value(
"pageserver_tenant_offloaded_timelines", {"tenant_id": f"{tenant_id}"}
)
assert offloaded_count == 0
ps_http.timeline_archival_config(
tenant_id,
leaf_timeline_id,
@@ -244,6 +249,11 @@ def test_timeline_offloading(neon_env_builder: NeonEnvBuilder, manual_offload: b
wait_until(leaf_offloaded)
wait_until(parent_offloaded)
offloaded_count = ps_http.get_metric_value(
"pageserver_tenant_offloaded_timelines", {"tenant_id": f"{tenant_id}"}
)
assert offloaded_count == 2
# Offloaded child timelines should still prevent deletion
with pytest.raises(
PageserverApiException,