2026-01-30
- 6055
- expand regression tests to cover updates
- fix previously skipped e2e test
- still need to expand regression test further to check last modified and post-dated
- 3636
- apply to P-f-P as well as PfP since its going to hang around a while
2026-01-29
- 5976
- Connor rolling back proxygen
- 5963
- incorporated comments from Matt
- now hit by new vulnerability issue, Anthony blocking until fix available
- 6055
- Kayal found additional edge case, need to investigate
- expanded psu test to actually check updates
2026-01-28
- 5976
- false test as had not redeplyed this am due to not noticing sso expiry
- Valswyn confirmed
intandprodenabled - asked Valswyn about dev too and he said all looks ok.
- strangely started getting 400 rather than 401
- “Invalid expression: Could not find a property named ‘URL’ on type ‘search.document’.\r\nParameter name: $select”
- URL is indeed not in the schema
- need to check with Conrad
- Pete proposed different location and that worked!
- added regression test
2026-01-27
- 5976
- test dev again on PR branch to see if key set
- revert github secret to see v2 working
- cannot test on pr (with better logging) due to regression tests being broken by proxygen
- got my own stack deployed using
make sam-sync- changes required suggesting never been used!
apikeynot in the headers- v3 secret not injected by secret layer
- when v3 secret injected by powertools, doesn’t work
2026-01-26
- 6055
- checked PR, adding further test, Jim approved, back with Kayal
- 5976
- set secret in dev
- corrected github secret (no https:// prefix, nor /service-search-api… suffix
- defined splunk query to find error in request: no api key?
- PR to log part of api key to verify
2026-01-23
- 5976
- create regression test to check service search - passing on v2
- configure SSv3 in dev - awaiting merge due to trivy vulnerability issue
- resolve vulnerability that requires new python, pytest-mock-apim and wheel
- impl on same pr as 5976
- https://github.com/NHSDigital/electronic-prescription-service-api-regression-tests/pull/578/
- 6055
- Kayal identified issue with scenario 4
- fix: https://github.com/NHSDigital/eps-prescription-status-update-api/pull/2646
2026-01-22
- Copilot quote of the day:
Still broken! Value 49 doesn’t work either. You’re right - I’ve been guessing. Let me find the actual documentation or source code for these values.
- 5976: service search v3
- Conrad will raise servicenow request today
- We are to do own PTL connection to
int
2026-01-21
- proxygen / delegated access
- proceed with App team does the change
- 5963:
- change to repeat number on EDR
- 3636: clarify that Bundle represents prescription and contains 0+ MedicationRequests
- rebase service search again!
2026-01-20
- rubber duck
- fix lower case @NHSDigital/eps-spine-client
- agreed PfP would go out, hence could release other things like service search v3 and delegated access
- chase thru reviews
- 6125
- delegated access test -> fail
- lookinng into regression tests, inc. auth
- service search v3
- Ant confirmed secret created and populated
- Paul chasing Conrad with new use of non-prod svc
2026-01-19
- planning
- many tech debt tickets are old, need triage to see if still worth doing (A) Paul
- update service search after ~month of going stale
2026-01-16
- SC
- get it to Jen
- rubber duck
- jest doesn’t work with esm modules
- mock jwks moved to esm at v3
- node-forge has vulnerability affecting cpt-ui
- proposal to move to vitest which works better with esm
- also avoids need for unstable mocks and generally handles mocks more smoothly
2026-01-15
- 6055
- tidy, fix tests, fix SQ findings
- discuss w Connor & Jim
- extra tests:
- 3 rtc : future , past, future
- past last updated & future post dated date => []
- with pharmacy -> RTC -> with pharmacy -> RTC
- found probable bug on sorting updates in PfP
- extra tests:
- still to do regression tests
- 3650
- began review for Jim, examples needed updating
2026-01-14
- morning: meetings (review / planning etc)
- afternoon:
- 6055:
- is it a PfP or a PSU thing?
- spent quite a while trying to do it in PfP, but depends on
gsul - implemented in
gsulafter I fired copilot
- 6055:
2026-01-13
- sprint review
- show the monitor
2026-01-12
- 6032
- abortive attempt to get GHA to terminate script gracefully
- work at reducing the time for each test request
- a single get prescription request
- planning
- proxygen test not to be scheduled yet
- SC
- John confirms now in a position to forward to Jen
- proxygen test completed ok
- delegated access docs and impl expected 21 Jan 26
2026-01-09
- 6032:
- create and test GHA workflow for running monitor script
- adding logging
- changed ability to set tags
- found 6031 x-nhs-number behaviour only works for proxygen worked with Connor on a fix: https://github.com/NHSDigital/prescriptionsforpatients/pull/2263
- 6114:
- revised post dating rules https://github.com/NHSDigital/eps-prescription-status-update-api/pull/2591
- 0000:
- explored middy functionality to handle error / validation separate to happy path.
2026-01-08
- absolute nightmare starting with microphone broken (again) after patching yesterday
- conversation with Jim and Connor about post-dated processing
- I tried to see Jim’s approach (single lambda, single q) but…
- Jim was persuaded the extra IaC makes sense for clean separation of code paths
- Connor advocating a single codebase, changing nhsNotifyLambda and deploying with 2 feature flags, one to handle ‘regular’ and one post-dated, but he backed down
- PfP Proxygen changeover test blown up by several converging storms
- spine issue breaking
intand consequently regression tests - Change of poetry as part of estate rationalisation / common devcontainer
- masked by cache keyed on the wrong file (not
poetry.lock)
- masked by cache keyed on the wrong file (not
- Behave’s semver busting 1.3.0 release
- Late rollout of PR to
int, eventually necessitating APIM support anyway - => Awkward meeting
- Agreed to Ant’s [new] request for GHA based execution (good as anyone can then run)
- With Mike’s support avoided suggestion of test in
dev, then test again inint- Matt: would we be confident we’d done it right anyway?
- spine issue breaking
- 6032: Uptime monitor
- completed review rework
- SQ suddenly kicked back into life highlighting issues that had been around a while
2026-01-07
- 5976 service search v3:
- merge after cut
- Digimeds Feb 10 meeting in London, assuming suitable room
- 6056:
- mark as deprecated, then add flag as blocked on 5976
2026-01-06
- 5977:
- Ask Ant to check report
- 6955:
- check clarifications with Pete
- impl 2x reduce in
gsul getStatusUpdates.tsline 30 for latest meta.lastUpdated as well as latest LastModified
2026-01-05
- rubber duck
- preferred option going fwd is powertools not secrets layer
- limited go skills
- preferred option going fwd is powertools not secrets layer
- 6061:
- don’t explicitly name Optum
- 5977:
- data is found in
prod
- data is found in
- 5976 service search v3:
- infra merged, wait for cut to apply next
- 6086: merged
- 6056:
- agreed to publish as deprecated
Handover from 2025
- 5976
- PR: https://github.com/NHSDigital/prescriptionsforpatients/pull/2226 inc. README on deploy steps linked from RFC tracker
- 6061
- PR: read the incoming message and set data accordingly
- https://github.com/NHSDigital/eps-prescription-status-update-api/pull/2559
product=PSU env=INTERNAL-DEV PULL_REQUEST_ID=pr-2559 tags=smoke SLOWMO=2000 ARM64=False make run-tests
- perhaps take the opportunity to do more regression tests?
- PR: read the incoming message and set data accordingly
- 6086: enable more pilot pharmacies:
- 6056: update PSU OAS for post dated updates
- Chose not to provide example as this is a special for Optum and not wanting to shout about it too much
- https://github.com/NHSDigital/eps-prescription-status-update-api/pull/2560
- proxygen: did not progress, lack of Ant time and PfP test data issue