fix(core): retry failed session wakes

packages/core/src/session/run-coordinator.ts

@@ -46,6 +46,7 @@ type Entry<A, E> = {
   interruptSeq?: number
   owner?: Fiber.Fiber<void, never>
   stopping: boolean
+  advisoryRetriesRemaining: number
 }
 
 /** Combines follow-up demand: runs dominate, while wakes retain the newest durable admission sequence. */
@@ -81,12 +82,17 @@ export const make = <Key, A, E>(options: {
       }),
     )
 
-    const makeEntry = (current: Demand, explicitWaiter?: Deferred.Deferred<A, E>): Entry<A, E> => ({
+    const makeEntry = (
+      current: Demand,
+      explicitWaiter?: Deferred.Deferred<A, E>,
+      advisoryRetriesRemaining = current._tag === "wake" ? 1 : 0,
+    ): Entry<A, E> => ({
       done: Deferred.makeUnsafe<A, E>(),
       settled: Deferred.makeUnsafe<Exit.Exit<A, E>>(),
       current,
       explicitWaiter,
       stopping: false,
+      advisoryRetriesRemaining,
     })
 
     const start = (key: Key, entry: Entry<A, E>, demand: Demand, successor = false) => {
@@ -132,6 +138,7 @@ export const make = <Key, A, E>(options: {
           const pending = entry.pending
           entry.pending = undefined
           entry.current = pending
+          entry.advisoryRetriesRemaining = pending._tag === "wake" ? 1 : 0
           start(key, entry, pending, true)
           return
         }
@@ -141,7 +148,12 @@ export const make = <Key, A, E>(options: {
         return
       }
 
-      const successor = entry.pending !== undefined ? makeEntry(entry.pending, entry.explicitWaiter) : undefined
+      const successor =
+        entry.pending !== undefined
+          ? makeEntry(entry.pending, entry.explicitWaiter)
+          : exit._tag === "Failure" && demand._tag === "wake" && !entry.stopping && entry.advisoryRetriesRemaining > 0
+            ? makeEntry(demand, entry.explicitWaiter, entry.advisoryRetriesRemaining - 1)
+            : undefined
       if (successor === undefined) active.delete(key)
       else active.set(key, successor)
       if (successor !== undefined) start(key, successor, successor.current, true)
@@ -151,6 +163,7 @@ export const make = <Key, A, E>(options: {
         exit._tag === "Failure" &&
         !(entry.stopping && Cause.hasInterruptsOnly(exit.cause)) &&
         demand._tag === "wake" &&
+        successor === undefined &&
         options.onFailure !== undefined
       ) {
         report(Effect.suspend(() => options.onFailure!(key, exit.cause)))

packages/core/src/session/runner/llm.ts

@@ -211,9 +211,10 @@ export const layer = Layer.effect(
       const model = yield* models.resolve(session)
       const entries = yield* SessionHistory.entriesForRunner(db, session.id, system.baselineSeq)
       const context = entries.map((entry) => entry.message)
+      const promptCacheKey = /^ses_[0-9a-f]{64}$/.test(session.id) ? session.id.slice(4) : session.id
       const request = LLM.request({
         model,
-        providerOptions: { openai: { promptCacheKey: session.id } },
+        providerOptions: { openai: { promptCacheKey } },
         system: [agent.info?.system, system.baseline]
           .filter((part): part is string => part !== undefined && part.length > 0)
           .map(SystemPart.make),

packages/core/test/session-run-coordinator.test.ts

@@ -33,7 +33,9 @@ describe("SessionRunCoordinator", () => {
     Effect.scoped(
       Effect.gen(function* () {
         const drained = yield* Deferred.make<void>()
-        const coordinator = yield* SessionRunCoordinator.make({ drain: () => Deferred.succeed(drained, undefined) })
+        const coordinator = yield* SessionRunCoordinator.make({
+          drain: () => Deferred.succeed(drained, undefined),
+        })
 
         yield* coordinator.wake("session")
         yield* Deferred.await(drained)
@@ -44,7 +46,9 @@ describe("SessionRunCoordinator", () => {
   it.effect("does nothing when interrupted while idle", () =>
     Effect.scoped(
       Effect.gen(function* () {
-        const coordinator = yield* SessionRunCoordinator.make({ drain: () => Effect.void })
+        const coordinator = yield* SessionRunCoordinator.make({
+          drain: () => Effect.void,
+        })
 
         yield* coordinator.interrupt("session")
       }),
@@ -55,7 +59,9 @@ describe("SessionRunCoordinator", () => {
     Effect.scoped(
       Effect.gen(function* () {
         let runs = 0
-        const coordinator = yield* SessionRunCoordinator.make({ drain: () => Effect.sync(() => runs++) })
+        const coordinator = yield* SessionRunCoordinator.make({
+          drain: () => Effect.sync(() => runs++),
+        })
 
         yield* coordinator.interrupt("session", 2)
         yield* coordinator.wake("session", 1)
@@ -722,6 +728,35 @@ describe("SessionRunCoordinator", () => {
     ),
   )
 
+  it.effect("retries a failed advisory successor once without another wake", () =>
+    Effect.scoped(
+      Effect.gen(function* () {
+        const firstGate = yield* Deferred.make<void>()
+        let runs = 0
+        const failure = new Error("transient wake failure")
+        const coordinator = yield* SessionRunCoordinator.make<string, void, Error>({
+          drain: () =>
+            Effect.sync(() => ++runs).pipe(
+              Effect.flatMap((run) =>
+                run === 1
+                  ? Deferred.await(firstGate).pipe(Effect.andThen(Effect.fail(failure)))
+                  : run === 2
+                    ? Effect.fail(failure)
+                    : Effect.void,
+              ),
+            ),
+        })
+
+        yield* coordinator.wake("session", 1)
+        yield* coordinator.wake("session", 2)
+        yield* Deferred.succeed(firstGate, undefined)
+        yield* coordinator.awaitIdle("session").pipe(Effect.exit)
+
+        expect(runs).toBe(3)
+      }),
+    ),
+  )
+
   it.effect("upgrades an active wake when an explicit run joins it", () =>
     Effect.scoped(
       Effect.gen(function* () {
@@ -916,8 +951,9 @@ describe("SessionRunCoordinator", () => {
         const failure = new Error("wake failed")
         const reported: Cause.Cause<Error>[] = []
         const reportedOnce = yield* Deferred.make<void>()
+        let runs = 0
         const coordinator = yield* SessionRunCoordinator.make<string, void, Error>({
-          drain: () => Effect.fail(failure),
+          drain: () => Effect.sync(() => runs++).pipe(Effect.andThen(Effect.fail(failure))),
           onFailure: (_key, cause) =>
             Effect.sync(() => reported.push(cause)).pipe(Effect.andThen(Deferred.succeed(reportedOnce, undefined))),
         })
@@ -927,6 +963,7 @@ describe("SessionRunCoordinator", () => {
         yield* Effect.yieldNow
 
         expect(reported).toHaveLength(1)
+        expect(runs).toBe(2)
         expect(Cause.squash(reported[0]!)).toBe(failure)
       }),
     ),

packages/core/test/session-runner.test.ts

@@ -2824,6 +2824,42 @@ describe("SessionRunnerLLM", () => {
     }),
   )
 
+  it.effect("bounds external session prompt cache keys", () =>
+    Effect.gen(function* () {
+      yield* setup
+      const externalSessionID = SessionV2.ID.fromExternal({
+        namespace: "discord",
+        key: "thread-one",
+      })
+      const otherExternalSessionID = SessionV2.ID.fromExternal({
+        namespace: "discord",
+        key: "thread-two",
+      })
+      yield* insertSession(externalSessionID)
+      yield* insertSession(otherExternalSessionID)
+      const session = yield* SessionV2.Service
+      yield* session.prompt({
+        sessionID: externalSessionID,
+        prompt: new Prompt({ text: "Run external session" }),
+        resume: false,
+      })
+      yield* session.prompt({
+        sessionID: otherExternalSessionID,
+        prompt: new Prompt({ text: "Run other external session" }),
+        resume: false,
+      })
+
+      requests.length = 0
+      yield* session.resume(externalSessionID)
+      yield* session.resume(otherExternalSessionID)
+
+      const keys = requests.map((request) => request.providerOptions?.openai?.promptCacheKey)
+      expect(keys).toEqual([externalSessionID.slice(4), otherExternalSessionID.slice(4)])
+      expect(keys.every((key) => typeof key === "string" && key.length === 64)).toBe(true)
+      expect(keys[0]).not.toBe(keys[1])
+    }),
+  )
+
   it.effect("fans out one failed run and allows a later retry", () =>
     Effect.gen(function* () {
       yield* setup