Closed randomir closed 1 year ago
Merging #540 (3d4d545) into master (d1bf2ee) will increase coverage by
0.16%
. The diff coverage is96.61%
.
@@ Coverage Diff @@
## master #540 +/- ##
==========================================
+ Coverage 86.69% 86.86% +0.16%
==========================================
Files 24 24
Lines 3473 3524 +51
==========================================
+ Hits 3011 3061 +50
- Misses 462 463 +1
Impacted Files | Coverage Δ | |
---|---|---|
dwave/cloud/api/client.py | 94.05% <93.75%> (+0.34%) |
:arrow_up: |
dwave/cloud/api/resources.py | 95.72% <95.83%> (-0.01%) |
:arrow_down: |
dwave/cloud/api/__init__.py | 100.00% <100.00%> (ø) |
|
dwave/cloud/api/constants.py | 100.00% <100.00%> (ø) |
|
dwave/cloud/api/models.py | 98.11% <100.00%> (+0.33%) |
:arrow_up: |
:mega: We’re building smart automated test selection to slash your CI/CD build times. Learn more
I believe Bulat might have some input for this feature. Can you add him to the PR as well?
We shouldn't release this until zero downtime for Leap is completed (Q3 '23). Otherwise, calls to /leap will end up failing during by weekly deployment. Bad UX.
We shouldn't release this until zero downtime for Leap is completed (Q3 '23). Otherwise, calls to /leap will end up failing during by weekly deployment. Bad UX.
@charleswhchan, Aren't SAPI calls failing during SAPI deploys?!
sapi-ws has zero downtime implemented and scheduler has a very short delay before restart, so all /sapi calls should be handled and processed as usual.
where as calls to /leap will show the maintenance page and probably return a 5xx error
We already use Leap API from the IDE, so I think it's fine to release this before the end of Q3. It's not like it will be visible and used by default. Currently it's not even documented; intended practically for internal use.
We shouldn't release this until zero downtime for Leap is completed (Q3 '23). Otherwise, calls to /leap will end up failing during by weekly deployment. Bad UX.
My 2c is that a user who is using this feature (especially given its current semi-private nature) would understand that when Leap is down they would get an error message. So IMO while it's not an ideal user experience, the relative benefits are well worth it.
My 2c is that a user who is using this feature (especially given its current semi-private nature) would understand that when Leap is down they would get an error message. So IMO while it's not an ideal user experience, the relative benefits are well worth it.
Rad, Bulat and myself will be meeting on Monday -- we can reach a decision then.
In this PR we add
LeapAPIClient
base client,LeapAccount
resource class, and the appropriate Leap API models (types) in order to support: