Currently all files are purged from imgix when replaced / updated / edited including skipped files.
Skipped files are not managed by imigix which leads to an error in the queue job.
Error: POST api/v1/purge returned 403: { "errors": [ { "detail": "Specified URL does not belong to your account.", "id": "60f6374266794d108a1b790d9c0e42ae", "meta": {}, "status": "403", "title": "Forbidden" } ], "jsonapi": { "version": "1.0" }, "meta": { "authentication": { "authorized": true, "clientId": null, "mode": "PUBLIC_APIKEY", "modeTitle": "Public API Key", "tag": "accounts@newism.com.au", "user": null }, "server": { "commit": "production", "status": { "healthy": true, "read_only": false, "tombstone": false }, "version": "3.213.15" } } }
In the case of a 403 (or 4xx) the job should probably complete successfully.
We could skip the job entirely if the file is skipped as a transform but I wonder what happens if there's a configuration change and a file that was processed by imgix is then skipped.
Currently all files are purged from imgix when replaced / updated / edited including skipped files.
Skipped files are not managed by imigix which leads to an error in the queue job.
In the case of a 403 (or 4xx) the job should probably complete successfully.
We could skip the job entirely if the file is skipped as a transform but I wonder what happens if there's a configuration change and a file that was processed by imgix is then skipped.