Book a Demo!
CoCalc Logo Icon
StoreFeaturesDocsShareSupportNewsAboutPoliciesSign UpSign In
gitpod-io
GitHub Repository: gitpod-io/gitpod
Path: blob/main/operations/observability/mixins/workspace/rules/central/ipfs.yaml
2506 views
1
# Copyright (c) 2023 Gitpod GmbH. All rights reserved.
2
# Licensed under the GNU Affero General Public License (AGPL).
3
# See License.AGPL.txt in the project root for license information.
4
5
apiVersion: monitoring.coreos.com/v1
6
kind: PrometheusRule
7
metadata:
8
labels:
9
prometheus: k8s
10
role: alert-rules
11
name: ipfs-monitoring-rules
12
spec:
13
groups:
14
- name: ipfs-alerts
15
rules:
16
- alert: IPFSStorageLow
17
labels:
18
severity: warning
19
team: engine
20
for: 10m
21
annotations:
22
runbook_url: https://github.com/gitpod-io/runbooks/blob/main/runbooks/IPFSAlmostOutOfStorage.md
23
summary: IPFS is running out of storage within a few days in a workspace cluster. Create and shift to a new workspace cluster with a fresh IPFS.
24
description: IPFS in cluster {{ $labels.cluster }} is running out of storage ({{ $value | humanize1024 }}B left). This happens naturally for workspace clusters over time, and is prevented by recreating the clusters regularly. Once out of storage, IPFS will stop working, and we fall back to the GCP registry for workspace image pulls, resulting in slower workspace startup times.
25
expr: min(node_filesystem_avail_bytes{device="/dev/mapper/lvm--disk-ipfs", node=~"services-.*", cluster!~"ephemeral.*"}) by (cluster) < 600 * 1024*1024*1024
26
- alert: IPFSStorageCritical
27
labels:
28
severity: critical
29
team: engine
30
for: 10m
31
annotations:
32
runbook_url: https://github.com/gitpod-io/runbooks/blob/main/runbooks/IPFSAlmostOutOfStorage.md
33
summary: IPFS is now very close to running out of storage in a workspace cluster. Create and shift to a new workspace cluster with a fresh IPFS.
34
description: IPFS in cluster {{ $labels.cluster }} has almost ran out of storage ({{ $value | humanize1024 }}B left). This happens naturally for workspace clusters over time, and is prevented by recreating the clusters regularly. Once out of storage, IPFS will stop working, and we fall back to the GCP registry for workspace image pulls, resulting in slower workspace startup times.
35
expr: min(node_filesystem_avail_bytes{device="/dev/mapper/lvm--disk-ipfs", node=~"services-.*", cluster!~"ephemeral.*"}) by (cluster) < 200 * 1024*1024*1024
36
37