Age | Commit message (Collapse) | Author | |
---|---|---|---|
2022-07-05 | Remove rate limits feature flags | feistel | |
Changelog: removed | |||
2022-02-21 | feat: Add TLS rate limits | Vladimir Shushlin | |
Changelog: added | |||
2022-01-11 | feat: add domain rate-limiter | Vladimir Shushlin | |
Changelog: added | |||
2021-10-14 | feat: add panic handler middleware | Jaime Martinez | |
Changelog: added | |||
2021-10-14 | feat: add source IP ratelimiter middleware | Jaime Martinez | |
It gets the source IP from `r.RemoteAddr` or from the `X-Forwarded-For` header for proxied requests (when `--listen-proxy` is enabled). The first iteration will only report logs and metrics when an IP is being rate limited. The rate limiter uses a Token Bucket approach using golang.org/x/time/rate, which can be configured with the newly added flags `rate-limit-source-ip` and `rate-limit-source-ip-burst`. To enable the rate limiter, set `rate-limit-source-ip` to value > 1, which is the number of requests per second to allow. It is enabled by default in "dry-run" mode so requests won't be dropped until the environment variable `FF_ENABLE_RATE_LIMITER` is set to `"true"`. See metrics.go for the newly added metrics. Changelog: added | |||
2021-08-26 | docs: fix method doc to pass code quality | feistel | |
2021-08-26 | refactor: remove metrics for disk source | feistel | |
2021-05-20 | Remove serverless code from Pages | feistel | |
2021-03-16 | Pass metrics into limiter to better decouple the instrumentation | Igor Wiedler | |
2021-03-16 | Instrument limit listener for saturation monitoring | Igor Wiedler | |
Changelog: added | |||
2020-12-22 | Reject all unknown http methods | Vladimir Shushlin | |
2020-10-13 | Update metrics | Kamil Trzciński | |
2020-10-13 | Split cache set/get and add metrics | Jaime Martinez | |
2020-10-01 | Move tracer to httptransport | Jaime Martinez | |
2020-10-01 | Rerun goimports | Jaime Martinez | |
2020-10-01 | Rename all zip and httprange metrics | Jaime Martinez | |
Renames all metrics to make them easier to understand. Adds a missing metric for number of open requests made by httprange. | |||
2020-09-30 | Add extra metric for currently cached files per archive | Jaime Martinez | |
2020-09-30 | Use histograms instead of gauges for transport durations | Jaime Martinez | |
2020-09-30 | Rename metrics and update types | Jaime Martinez | |
2020-09-30 | Add tests for new metrics | Jaime Martinez | |
DRY transport. Add buckets to responsiveness metric. | |||
2020-09-30 | Add more metrics for zip serving | Jaime Martinez | |
Adds a bunch of new metrics related to https://gitlab.com/gitlab-org/gitlab-pages/-/issues/423. It uses [httptrace.ClienTrace](https://golang.org/src/net/http/httptrace/trace.go) to add a bunch of very granular metrics that happen when an http connection is established. | |||
2020-09-25 | Simplify after code review | Kamil Trzciński | |
2020-09-24 | Add Name to the VFS interface to use with metrics | Jaime Martinez | |
2020-09-24 | Rebase from base branch | Jaime Martinez | |
Add vfs.VFS implementation for zip Fix build errors Clean zip vfs Add tests for Root Add zip serving metric Return a zip.Instance() when source == zip Add simple acceptance test for zip serving Use correct contents No need to start testServer in go routine | |||
2020-09-15 | Apply suggestions from feedback | Jaime Martinez | |
Update metric name | |||
2020-09-15 | Move setup to helper function | Jaime Martinez | |
2020-09-15 | Update metrics acceptance tests | Jaime Martinez | |
Remove metrics_test.go and leave responsibility of testing to acceptance_test.go with `TestPrometheusMetricsCanBeScraped`. Update TODOs with issue links. | |||
2020-09-15 | Add metrics for httprange and trace logs | Jaime Martinez | |
Add metrics and trace logs Enables meteredRoundTripper to log response data in a Traceln() call. It adds a name to the round tripper for easy logging. Adds two new metrics to be used by the `httprange` package to track number of requests made to object storage. Apply 1 suggestion(s) to 1 file(s) (cherry picked from commit 57a98108c728ba6e6575a95899455d5fac536ae1) | |||
2020-08-13 | Add VFS for local disk | Jacob Vosmaer | |
2020-06-11 | Add units to metric names | Jaime Martinez | |
2020-06-10 | Add buckets for histograms | Jaime Martinez | |
2020-06-09 | Use histogram without labels | Jaime Martinez | |
Register ServingTime. Rename DiskServingFileSize so that it's a histogram without needing labels | |||
2020-06-05 | Use histograms instead of gagues | Jaime Martinez | |
Removed test assertion for metric as histograms are not collectors so we can't really test it. Left the test though as it's still valuable. | |||
2020-06-04 | Add serving time metric for all files | Jaime Martinez | |
2020-06-03 | Add file size metric for disk serving | Jaime Martinez | |
2020-05-04 | Add a prometheus counter for GitLab API call failures | Grzegorz Bizon | |
2020-02-11 | Add prometheus metrics for GitLab API client | Jaime Martinez | |
Refactor metrics initialization removing init function from the metrics package. | |||
2020-02-03 | Apply suggestion to metrics/metrics.go | Grzegorz Bizon | |
2020-02-03 | Add serverless serving metrics | Grzegorz Bizon | |
This adds: - serverless invocations counter - serverless request latency histogram | |||
2020-01-27 | Add domain configuration duration (from disk) to Prometheus metrics | Krasimir Angelov | |
2019-12-17 | Add GitLab API domains source caching metrics | Grzegorz Bizon | |
2019-08-30 | Add full HTTP metrics and logging to GitLab pages using LabKit | Andrew Newdigate | |
2019-08-08 | Update metrics names. | Ben Kochie | |
* Rename "DomainsServed", `_total` is reserved for counters. * Rename "SessionsActive", "active sessions" makes more sense. | |||
2018-08-15 | Make it clear in commen tthat DomainUpdates pertains to successful updates | Stan Hu | |
2018-08-15 | Make it explicit that when ReadGroups fails initially, an error is logged | Stan Hu | |
2018-08-14 | Abort domain scan if a failure is encountered | Stan Hu | |
This prevents the total domain list to be cleared out completely while the system is running. Closes https://gitlab.com/gitlab-com/infrastructure/issues/4749 | |||
2017-09-22 | Fix some lint failures | Nick Thomas | |
2017-02-07 | Improve complexity metrics | Z.J. van de Weg | |
Also fixed the dependencies, moved metrics to its own package and updated the tests |