artsy / force

The Artsy.net website
https://www.artsy.net
MIT License
599 stars 152 forks source link

feat: propagate client-side-cached-but-not-server custom 0 TTL to CDN #14434

Closed mzikherman closed 2 months ago

mzikherman commented 2 months ago

Untested! But saw the conversation in Slack.

relativeci[bot] commented 2 months ago

#180 Bundle Size — 9.54MiB (+0.01%).

1381368(current) vs 96ca453 main#178(baseline)

[!WARNING] Bundle contains 45 duplicate packages – View duplicate packages

Bundle metrics  Change 2 changes Regression 1 regression
                 Current
#180
     Baseline
#178
Regression  Initial JS 4.32MiB(+0.03%) 4.32MiB
No change  Initial CSS 0B 0B
Change  Cache Invalidation 0.86% 42.38%
No change  Chunks 142 142
No change  Assets 145 145
No change  Modules 5552 5552
No change  Duplicate Modules 380 380
No change  Duplicate Code 4.78% 4.78%
No change  Packages 292 292
No change  Duplicate Packages 42 42

Bundle size by type  Change 1 change Regression 1 regression
|            |       Current
[#180](https://app.relative-ci.com/projects/zNeNymVCPwxxOLTcM9qt/jobs/180-FBildXWaQ5lTAnsbhI7e?utm_source=github&utm_medium=pr-report "View bundle analysis report") |      Baseline
[#178](https://app.relative-ci.com/projects/zNeNymVCPwxxOLTcM9qt/jobs/178-xC0S6DGOimUosjXEMFCL?utm_source=github&utm_medium=pr-report "View baseline bundle analysis report") | |:--|--:|--:| | Regression  [JS](https://app.relative-ci.com/projects/zNeNymVCPwxxOLTcM9qt/jobs/180-FBildXWaQ5lTAnsbhI7e/assets?ba=%7B%22filters%22%3A%22ft.CSS-0_ft.JS-1_ft.IMG-0_ft.MEDIA-0_ft.FONT-0_ft.HTML-0_ft.OTHER-0%22%7D "View JS assets") | `9.33MiB` (`+0.01%`) | `9.33MiB` | | No change  [Other](https://app.relative-ci.com/projects/zNeNymVCPwxxOLTcM9qt/jobs/180-FBildXWaQ5lTAnsbhI7e/assets?ba=%7B%22filters%22%3A%22ft.CSS-0_ft.JS-0_ft.IMG-0_ft.MEDIA-0_ft.FONT-0_ft.HTML-0_ft.OTHER-1%22%7D "View Other assets") | `213.94KiB` | `213.94KiB` |

Bundle analysis reportBranch propagate-zero-ttlProject dashboard


Generated by RelativeCIDocumentationReport issue

damassi commented 2 months ago

@joeyAghion - this behaves the same as before, with redis: with serverCacheTTL: 0 the desired behavior was indeed to apply to any query keyed by that route to not cache -- because we don't want things to cache on those routes! So this seems correct (or rather, correct in the sense of the original design).