fix(inference-gateway): backport EPP request-body forwarding (#11991) and prefill reservations (#13042) to v1.4.2 - #2

Merged

avinash-rafay merged 15 commits into

v1.4.2-fixesfrom

cherry-pick/v1.4.2-fixes/11991-13042

Sep 1, 2026

Commits on Sep 1, 2026

fix(inference-gateway): forward full request body to EPP KV-router scorer (#11991)

avinash-rafay
and
avinashpenmetsa
committed
fe94c91
fix(epp): reserve prefill load during routing

tmonty12
authored and
avinashpenmetsa
committed
4d54129
refactor(epp): remove advisory prefill route

tmonty12
authored and
avinashpenmetsa
committed
025b97b
fix(epp): cancel pending prefill reservations

tmonty12
authored and
avinashpenmetsa
committed
84e8f29
fix(epp): bound prefill reservation lifecycle

tmonty12
authored and
avinashpenmetsa
committed
ea7e01d
refactor(epp): scope reservations to C handle

tmonty12
authored and
avinashpenmetsa
committed
260b305
fix(epp): harden booking cleanup

tmonty12
authored and
avinashpenmetsa
committed
07a59fc
fix(epp): harden reservation ownership

tmonty12
authored and
avinashpenmetsa
committed
f0e647e
fix(epp): bound booking cleanup executor

tmonty12
authored and
avinashpenmetsa
committed
054e053
fix(epp): release late reservation errors

tmonty12
authored and
avinashpenmetsa
committed
f2b062d
fix(router): bound stale reservation reaping

tmonty12
authored and
avinashpenmetsa
committed
08e3fc9
fix(epp): raise prefill admission defaults

tmonty12
authored and
avinashpenmetsa
committed
3e8b2f9
Potential fix for pull request finding 'CodeQL / Incorrect conversion between integer types'

authored and
avinashpenmetsa
committed
0284de3
fix(inference-gateway): adapt prefill reservation port to 1.4.2 APIs
avinashpenmetsa
committed
65c3194
fix(inference-gateway): guard nil prefill cancellation hook
avinashpenmetsa
committed
0a12bfb

Navigation Menu

Uh oh!

fix(inference-gateway): backport EPP request-body forwarding (#11991) and prefill reservations (#13042) to v1.4.2 - #2

fix(inference-gateway): backport EPP request-body forwarding (#11991) and prefill reservations (#13042) to v1.4.2#2
avinash-rafay merged 15 commits into
v1.4.2-fixesfrom
cherry-pick/v1.4.2-fixes/11991-13042

Commits on Sep 1, 2026

fix(inference-gateway): forward full request body to EPP KV-router scorer (#11991)

fix(epp): reserve prefill load during routing

refactor(epp): remove advisory prefill route

fix(epp): cancel pending prefill reservations

fix(epp): bound prefill reservation lifecycle

refactor(epp): scope reservations to C handle

fix(epp): harden booking cleanup

fix(epp): harden reservation ownership

fix(epp): bound booking cleanup executor

fix(epp): release late reservation errors

fix(router): bound stale reservation reaping

fix(epp): raise prefill admission defaults

Potential fix for pull request finding 'CodeQL / Incorrect conversion between integer types'

fix(inference-gateway): adapt prefill reservation port to 1.4.2 APIs

fix(inference-gateway): guard nil prefill cancellation hook