In this CI job, we have a query with a timeout of 180 seconds that fails its first try after 440 seconds. I would think that it makes sense to limit the duration of a try by default, with some time longer than the [timeout:] setting. Currently you could only do that with run_timeout_secs, but that is intended to be a timeout for all tries of a query; not a single one. With [timeout:180] you probably want a run_timeout_secs that is a multiple of that 180 seconds, but less than that for individual tries.
This is now possible using RequestTimeout. The default is an additional 20 seconds on top of the query [timeout:], so a request with [timeout:180] would timeout after 200 seconds.
In this CI job, we have a query with a timeout of 180 seconds that fails its first try after 440 seconds. I would think that it makes sense to limit the duration of a try by default, with some time longer than the
[timeout:]
setting. Currently you could only do that withrun_timeout_secs
, but that is intended to be a timeout for all tries of a query; not a single one. With[timeout:180]
you probably want arun_timeout_secs
that is a multiple of that 180 seconds, but less than that for individual tries.