fix: add concurrency controls for multiple sessions using the same provider (#1069)

* fix: add concurrency controls for multiple sessions using the same provider

Adds OS-inspired scheduling primitives (from HiveMind/AIMD research) to prevent
concurrent sessions from the same provider from experiencing slowdowns, random
stops, and cascading failures.

Server-side:
- Health check skips OpenCode restart when sessions are actively busy — a busy
  server under concurrent load can fail the health check timeout without being
  dead. Staleness guard forces restart if unhealthy+busy persists >2 minutes.
- Upstream SSE stall timeout scaled from 20s to 60s to avoid unnecessary
  reconnections when multiple sessions are waiting for LLM responses.

Client-side (HiveMind primitives, arXiv:2604.17111):
- Transparent retry with exponential backoff (1s→2s→4s, max 32s) for
  429/502/503/504 errors — the #1 most effective primitive from the paper.
- Circuit breaker: opens after 3 consecutive retryable errors, cooldown
  doubles each trip (30s→60s→120s, capped 128s), matching TCP AIMD.
- Per-provider session tracking with TTL eviction (1h idle sweep).
- Fetch-level retry gated on AbortError/TypeError only (not DNS failures).

Refs github-code-review skill findings (all 8 issues resolved).

* fix: use definite assignment assertion for response variable

Fixes TS2454: Variable 'response' is used before being assigned
in strict mode. The for-loop body always assigns it on every path
that reaches the post-loop code, but TS can't prove that.

* fix: add cleanupSession to error paths and remove unreachable code

P1 fixes (Greptile review):
- cleanupSession called on fetch error throw path
- cleanupSession called on non-retryable HTTP error throw path
- Removed unreachable post-loop code (loop always terminates via return or throw)

Adds explicit post-loop throw to satisfy TypeScript strict return check.

* fix: address Greptile review feedback on concurrent session controls

Removes client-side session tracking that leaked on normal completion paths.

The session tracking was redundant — the server-side health check already reads from

sessionRuntime.getSessionActivitySnapshot() for busy-session detection.

Changes:

- Remove activeSessions Set and all session-tracking functions from provider-tracker

- Remove trackSessionStarted/cleanupSession calls from client.ts

- Remove unreachable (response as Response) block after retry loop

- Make upstreamStallTimeoutMs conditional: 60s when >1 sessions, 20s otherwise

Refs #1069

* fix: enforce dynamic concurrency safeguards

---------

Co-authored-by: Bohdan Triapitsyn <artmore@protonmail.com>

packages/ui/src/lib/opencode/provider-tracker.ts

@@ -0,0 +1,133 @@
+/**
+ * Provider Circuit-Breaker & Retry Tracker
+ *
+ * Tracks per-provider error state to enable:
+ * - Transparent retry with exponential backoff for transient errors
+ * - Circuit breaking (pause requests to a provider during error storms)
+ *
+ * Inspired by HiveMind (arXiv:2604.17111) OS-inspired scheduling primitives.
+ */
+
+const DEFAULT_CIRCUIT_BREAK_THRESHOLD = 3
+const DEFAULT_CIRCUIT_COOLDOWN_MS = 30_000
+const DEFAULT_RETRY_BASE_DELAY_MS = 1000
+const DEFAULT_RETRY_MAX_DELAY_MS = 32_000
+const DEFAULT_RETRY_MAX_ATTEMPTS = 3
+const PROVIDER_EVICTION_TTL_MS = 60 * 60 * 1000
+const PROVIDER_EVICTION_INTERVAL_MS = 10 * 60 * 1000
+
+const RETRYABLE_STATUS_CODES = new Set([429, 502, 503, 504])
+
+type ProviderState = {
+  consecutiveErrors: number
+  lastErrorAt: number
+  circuitOpen: boolean
+  circuitOpenAt: number
+  circuitCooldownMs: number
+}
+
+const providers = new Map<string, ProviderState>()
+
+function evictStaleProviders(): void {
+  const now = Date.now()
+  for (const [providerID, state] of providers) {
+    if (
+      state.consecutiveErrors === 0 &&
+      now - state.lastErrorAt > PROVIDER_EVICTION_TTL_MS
+    ) {
+      providers.delete(providerID)
+    }
+  }
+}
+
+if (typeof setInterval !== 'undefined') {
+  setInterval(evictStaleProviders, PROVIDER_EVICTION_INTERVAL_MS)
+}
+
+function getOrCreateProvider(providerID: string): ProviderState {
+  let state = providers.get(providerID)
+  if (!state) {
+    state = {
+      consecutiveErrors: 0,
+      lastErrorAt: 0,
+      circuitOpen: false,
+      circuitOpenAt: 0,
+      circuitCooldownMs: DEFAULT_CIRCUIT_COOLDOWN_MS,
+    }
+    providers.set(providerID, state)
+  }
+  return state
+}
+
+export function recordProviderSuccess(providerID: string): void {
+  if (!providerID) return
+  const state = providers.get(providerID)
+  if (!state) return
+  state.consecutiveErrors = 0
+  state.lastErrorAt = 0
+}
+
+export function recordProviderError(providerID: string, status?: number): void {
+  if (!providerID) return
+  const state = getOrCreateProvider(providerID)
+  state.consecutiveErrors += 1
+  state.lastErrorAt = Date.now()
+
+  if (
+    isCircuitBreakerStatus(status) &&
+    state.consecutiveErrors >= DEFAULT_CIRCUIT_BREAK_THRESHOLD
+  ) {
+    state.circuitOpen = true
+    state.circuitOpenAt = Date.now()
+    console.warn(
+      `[provider-tracker] Circuit opened for ${providerID} after ${state.consecutiveErrors} consecutive errors`
+    )
+  }
+}
+
+function isCircuitBreakerStatus(status?: number): boolean {
+  return status !== undefined && RETRYABLE_STATUS_CODES.has(status)
+}
+
+export function isCircuitOpen(providerID: string): boolean {
+  const state = providers.get(providerID)
+  if (!state?.circuitOpen) return false
+
+  const elapsed = Date.now() - state.circuitOpenAt
+  if (elapsed >= state.circuitCooldownMs) {
+    state.circuitOpen = false
+    state.consecutiveErrors = 0
+    state.circuitCooldownMs = Math.min(
+      state.circuitCooldownMs * 2,
+      DEFAULT_RETRY_MAX_DELAY_MS * 4
+    )
+    return false
+  }
+
+  return true
+}
+
+export function shouldRetry(providerID: string, status: number, attempt: number): boolean {
+  if (!RETRYABLE_STATUS_CODES.has(status)) return false
+  if (attempt >= DEFAULT_RETRY_MAX_ATTEMPTS - 1) return false
+  if (isCircuitOpen(providerID)) return false
+  return true
+}
+
+export function assertProviderCircuitClosed(providerID: string): void {
+  if (!providerID || !isCircuitOpen(providerID)) return
+  throw new Error(`Provider ${providerID} is temporarily unavailable after repeated errors. Please retry shortly.`)
+}
+
+export function getRetryDelayMs(attempt: number): number {
+  const delay = DEFAULT_RETRY_BASE_DELAY_MS * 2 ** attempt
+  return Math.min(delay, DEFAULT_RETRY_MAX_DELAY_MS)
+}
+
+export function resetCircuit(providerID: string): void {
+  const state = providers.get(providerID)
+  if (!state) return
+  state.consecutiveErrors = 0
+  state.circuitOpen = false
+  state.circuitCooldownMs = DEFAULT_CIRCUIT_COOLDOWN_MS
+}