FlexMeasures / flexmeasures-client / 42011499910
96%

DEFAULT BRANCH: main
Ran
Jobs 10
Files 6
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 95.681%. Remained the same
42011499910

push

github

web-flow 
feat(FRBC): ask again for a schedule that failed to arrive (#254)

* feat(FRBC): ask again for a schedule that failed to arrive

A scheduling job can fail on a transient of its own: a worker crashing out of its solver, or a
pandas error over the windows two of these twelve-hour schedules overlap in. FlexMeasures also
refuses a trigger outright where the state of charge a schedule would be planned from has not
landed yet, which is a race this handler can lose at startup.

Unanswered, any of those costs the storage its instruction for that window. The resource manager
is told nothing and goes on polling, and nothing else notices either, since asking for a schedule
runs as a task whose exception is logged and dropped. One such failure held a co-simulation open
for the rest of its run: a worker took SIGABRT inside HiGHS, and the resource manager behind it
polled until its round timed out, every round.

So a schedule is now asked for up to schedule_attempts times, two seconds apart, both
configurable, with one attempt meaning the handler behaves as it did before.

A cause that fails identically however often it is asked is never asked about again, since asking
only multiplies failed jobs over the server's workers: a scheduler colliding with its own belief,
and an infeasible problem. Infeasibility reaches this path at all because FlexMeasures retired its
fallback scheduler in v1.0.0 and fails the job itself now. A trigger refused outright is not asked
again either, bar the one refusal that answers a race: a flex model the server will not take is
not one it will take two seconds later.

Every attempt says what is known when it is made, rather than carrying the first attempt's prior.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CcpawsJs74cSwFLXkz7ZwW
Signed-off-by: F.N. Claessen <claessen@seita.nl>

* feat(FRBC): say so where a fallback scheduler answered instead

FlexMeasures answered an infeasible p... (continued)

997 of 1042 relevant lines covered (95.68%)

9.57 hits per line

Jobs
ID Job ID Ran Files Coverage
1 ubuntu-latest - py3.12 - 42011499910.1 6
95.68
GitHub Action Run
2 macos-latest - py3.12 - 42011499910.2 6
95.68
GitHub Action Run
3 macos-latest - py3.14 - 42011499910.3 6
95.68
GitHub Action Run
4 macos-latest - py3.13 - 42011499910.4 6
95.68
GitHub Action Run
5 ubuntu-latest - py3.13 - 42011499910.5 6
95.68
GitHub Action Run
6 macos-latest - py3.11 - 42011499910.6 6
95.68
GitHub Action Run
7 ubuntu-latest - py3.11 - 42011499910.7 6
95.68
GitHub Action Run
8 ubuntu-latest - py3.10 - 42011499910.8 6
95.68
GitHub Action Run
9 ubuntu-latest - py3.14 - 42011499910.9 6
95.68
GitHub Action Run
10 macos-latest - py3.10 - 42011499910.10 6
95.68
GitHub Action Run
Source Files on build 42011499910