| Commit message (Collapse) | Author | Age | Lines | ||
|---|---|---|---|---|---|
| ... | |||||
| * | Update Sir Robin to CJ11 (#399) |  Boris Muratov | 2024-07-03 | -1/+1 | |
| | | |||||
| * | Move noqa definition required in latest ruff version |  Chris Lovering | 2024-07-01 | -2/+2 | |
| | | |||||
| * | Allow new kube-state-metrics image to watch ingresses |  Joe Banks | 2024-07-01 | -0/+1 | |
| | | |||||
| * | Move away from vendored kube-state-metrics | Joe Banks | 2024-07-01 | -1/+1 | |
| | | |||||
| * | Add issuer for Vault certificates in tooling namespace | Joe Banks | 2024-06-27 | -0/+5 | |
| | | | | | | We will use this to deploy internal TLS certificates from a self-signed CA that allows for TLS traffic within the cluster. | ||||
| * | Add deployment of Keycloak | Joe Banks | 2024-06-27 | -0/+122 | |
| | | |||||
| * | Scale AM back to 3 replicas | Chris Lovering | 2024-06-24 | -1/+1 | |
| | | |||||
| * | Add ff-bot deployment | Joe Banks | 2024-06-16 | -0/+82 | |
| | | |||||
| * | Add Kubernetes volume alerts | Joe Banks | 2024-06-16 | -0/+11 | |
| | | | | | | | | | | | | It seems that Linode has added storage reporting info to the CSI driver allowing us to pick up on the storage use of persistent volume claims within the cluster. This creates and deploys an alert that will report if any volume has under 10% of space left. I have excluded Prometheus as our TSDB retention settings mean that it will always stay just below it's volume size by design. | ||||
| * | Update Loki config with new compactor preferences for retention modes | Joe Banks | 2024-06-13 | -1/+6 | |
| | | | | | | | | | | * `retention_enabled`: enable retention mode within the compactor * `delete_request_store`: store deletion requests within the s3 cluster that is also used to house log chunks * `delete_request_cancel_period`: do not exercise log deletion instructions until at least one hour has passed to prevent accidental deletion | ||||
| * | Update Prometheus deployment with a tmpfs for the reloader | Joe Banks | 2024-06-10 | -0/+9 | |
| | | |||||
| * | Add secrets for reloader webhook | Joe Banks | 2024-06-10 | -0/+0 | |
| | | |||||
| * | Add sidecar container to reload Prometheus config on change | Joe Banks | 2024-06-10 | -0/+25 | |
| | | |||||
| * | Add reloader hook configmap to reload prometheus on change | Joe Banks | 2024-06-10 | -0/+38 | |
| | | |||||
| * | Add Alert for Prometheus config reload failure | Joe Banks | 2024-06-10 | -0/+9 | |
| | | |||||
| * | Enable scraping of Prometheus pods | Joe Banks | 2024-06-10 | -0/+3 | |
| | | |||||
| * | Update Pinnwand logo to square image | Joe Banks | 2024-06-09 | -1/+1 | |
| | | |||||
| * | Update from command to args in site deployment | Joe Banks | 2024-06-07 | -1/+1 | |
| | | | | | | | | | Kubernetes renames ENTRYPOINT in Docker images to command and any additional args go in `args` (confusing, I know!) This ensures that we run within the context of Poetry so can reach Django and other installed requirements when running migrations. | ||||
| * | Remove unnecessary shell execution for migration initContainer | Joe Banks | 2024-06-07 | -3/+3 | |
| | | |||||
| * | Update site to run migrations in an init container | Joe Banks | 2024-06-07 | -0/+13 | |
| | | | | | | | | | | | | | | | | | | | In accordance with updates from python-discord/site#1338 this changes the way migrations are run. Previously, migrations would be run all from within the manage.py execution process with the command being manually spawned using Django internals. After python-discord/site#1338 merges the Dockerfile will directly invoke gunicorn and bypass manage.py to simplify the process and avoid problems with shared database contexts. Hence, we need to manually run migrations using an init container. With testing there is no additional delay in doing this as spinning up an init container is cheap and we don't cut over any traffic until the site passes a healthcheck anyway. | ||||
| * | Rename relabelledpods to just pods | Joe Banks | 2024-06-07 | -1/+1 | |
| | | | | | | | | This was a redundant rename and reduced the clarity of jobs when querying from inside Grafana. This rectifies that by renaming the stream to just `pods`. | ||||
| * | Reflect pydis.wtf certificate into Loki namespace | Joe Banks | 2024-06-07 | -2/+2 | |
| | | |||||
| * | Add secret for Loki authentication | Joe Banks | 2024-06-07 | -0/+0 | |
| | | |||||
| * | Add new Ingress for Loki gateway | Joe Banks | 2024-06-07 | -0/+25 | |
| | | |||||
| * | Add Metricity manifest | Joe Banks | 2024-06-06 | -0/+30 | |
| | | | | | Copies the Metricity deployment manifest from the Metricity repo. | ||||
| * | Add tmpfs to King Arthur | Joe Banks | 2024-06-05 | -0/+9 | |
| | | |||||
| * | Remove PostgreSQL Exporter from Kubernetes | Joe Banks | 2024-06-02 | -55/+0 | |
| | | |||||
| * | Remove Kubernetes PostgreSQL Alerts | Joe Banks | 2024-06-02 | -29/+0 | |
| | | |||||
| * | Remove Kubernetes PostgreSQL backup from Blackbox | Joe Banks | 2024-06-02 | -6/+1 | |
| | | |||||
| * | Remove PostgreSQL deployment from Kubernetes | Joe Banks | 2024-06-02 | -127/+0 | |
| | | |||||
| * | Update pixels environment variable | Joe Banks | 2024-06-02 | -0/+0 | |
| | | |||||
| * | Update Metabase configuration secret | Joe Banks | 2024-06-02 | -0/+0 | |
| | | |||||
| * | Update site secret with new database address | Joe Banks | 2024-06-01 | -0/+0 | |
| | | |||||
| * | Update site and metricity with new metricity db user credentials | Joe Banks | 2024-05-28 | -0/+0 | |
| | | |||||
| * | Update kube-system namespace docs with new metrics-server details | Joe Banks | 2024-05-28 | -4/+5 | |
| | | |||||
| * | Add Helm deployment info for metrics-server | Joe Banks | 2024-05-28 | -0/+24 | |
| | | | | | | | | Due to the way Linode seems to issue certificates for our nodes, we need to disable TLS verification for communications to fetch metric information. It's unfortunate but non-critical and it does restore metrics-server functionality. | ||||
| * | Add documentation on services deployed to the kube-system namespace | Joe Banks | 2024-05-28 | -0/+33 | |
| | | |||||
| * | Add new ServiceAccount for cert issuance | Joe Banks | 2024-05-27 | -0/+5 | |
| | | |||||
| * | Update mTLS bundle for ingress-nginx | Joe Banks | 2024-05-27 | -36/+46 | |
| | | |||||
| * | Add Helm instructions for Vault | Joe Banks | 2024-05-27 | -0/+54 | |
| | | |||||
| * | Add pydis.wtf cert to vault namespace | Joe Banks | 2024-05-27 | -2/+2 | |
| | | |||||
| * | Fix AlertManager Discord instance formatting | Joe Banks | 2024-05-27 | -1/+1 | |
| | | | | | | | | | | | | We made a change to include the instance in alerts sent to Discord, but not all of our configured alerts send this field. As a result, we would have incorrectly formatted alerts being sent through to Discord which were tricky to read. The format template has now been changed to only conditionally render the instance label if it is present on a triggered alert. | ||||
| * | Take 15 minutes before alerting on high latency |  Johannes Christ | 2024-05-20 | -2/+2 | |
| | | |||||
| * | Instruct code jam management to connect to lovelace | Johannes Christ | 2024-05-18 | -0/+0 | |
| | | |||||
| * | Instruct black knight to connect to lovelace | Johannes Christ | 2024-05-18 | -0/+0 | |
| | | |||||
| * | Annotations.instance => Labels.instance | Joe Banks | 2024-05-18 | -1/+1 | |
| | | |||||
| * | Add instance to AlertManager Discord embeds | Joe Banks | 2024-05-17 | -1/+1 | |
| | | |||||
| * | Update Bitwarden Kubernetes secret with new database location | Joe Banks | 2024-05-17 | -0/+0 | |
| | | |||||
| * | Bump limits and requests for bots that have been OOMing recently | Chris Lovering | 2024-05-16 | -3/+3 | |
| | | |||||
| * | Move AlertManager to 4 replicas | Joe Banks | 2024-05-16 | -1/+1 | |
| | | |||||
 |
index : infra | |
| The PyDis DevOps boiler house, the engine room, where the magic happens. |
| aboutsummaryrefslogtreecommitdiffstats |