feat(run): durable checkpoint + resume (wire Ports.Checkpointer) #20

Merged
steve merged 2 commits from feat/kernel-checkpoint into main 2026-06-29 20:44:17 +00:00
7 changed files with 98 additions and 39 deletions
Showing only changes of commit 38d656ec71 - Show all commits
+1 -2
View File
@@ -35,10 +35,9 @@ type RunCheckpointMeta struct {
// RunCheckpoint is one persisted snapshot of a run's resumable progress.
type RunCheckpoint struct {
Meta RunCheckpointMeta
Messages []llm.Message // conversation so far (single-loop or active phase)
Messages []llm.Message // conversation so far (single-loop runs)
Iteration int // completed agent-loop iterations
CompletedPhases []run.PhaseOutput // finished phases, in order (multi-phase agents)
ActivePhase string // current phase name (multi-phase agents); "" otherwise
UpdatedAt time.Time
}
+1 -3
View File
@@ -58,7 +58,6 @@ func (h *handle) Save(ctx context.Context, st run.RunCheckpointState) error {
Messages: st.Messages,
Iteration: st.Iteration,
CompletedPhases: st.CompletedPhases,
ActivePhase: st.ActivePhase,
UpdatedAt: now,
}); err != nil {
return err
1
@@ -92,7 +91,6 @@ func (noop) Fail(context.Context, error) error { return nil }
type factory struct {
store CheckpointStore
throttle time.Duration
now func() time.Time
}
Review

factory.now field is never assigned (NewFactory omits it) — dead/always-nil clock field

maintainability · flagged by 1 model

  • checkpoint/handle.go:95factory.now is dead. The factory struct carries a now func() time.Time field, but NewFactory never sets it (handle.go:103-105), so f.now is always nil and Begin always passes nil to New (handle.go:122), which then defaults to time.Now. Unlike handle, there's no constructor/option that injects a clock, so the field can never be anything but nil. Either drop it, or add the injection path it implies (mirroring handle's testable clock).

🪰 Gadfly · advisory

⚪ **factory.now field is never assigned (NewFactory omits it) — dead/always-nil clock field** _maintainability · flagged by 1 model_ - **`checkpoint/handle.go:95` — `factory.now` is dead.** The `factory` struct carries a `now func() time.Time` field, but `NewFactory` never sets it (handle.go:103-105), so `f.now` is always nil and `Begin` always passes nil to `New` (handle.go:122), which then defaults to `time.Now`. Unlike `handle`, there's no constructor/option that injects a clock, so the field can never be anything but nil. Either drop it, or add the injection path it implies (mirroring `handle`'s testable clock). <sub>🪰 Gadfly · advisory</sub>
var _ run.CheckpointerFactory = (*factory)(nil)
2
@@ -119,5 +117,5 @@ func (f *factory) Begin(_ context.Context, info run.RunInfo) (run.Checkpointer,
ModelTier: info.ModelTier,
ParentRunID: info.ParentRunID,
}
return New(f.store, meta, f.throttle, f.now), nil
return New(f.store, meta, f.throttle, nil /* now defaults to time.Now */), nil
}
+15 -6
View File
@@ -3,6 +3,7 @@ package run
import (
"context"
"errors"
"log/slog"
"gitea.stevedudenhoeffer.com/steve/majordomo/llm"
)
@@ -17,12 +18,14 @@ import (
// ResumeState carries a recovered run's prior progress into Run so the run
// continues instead of restarting. The host's recovery path sets it via
// WithResumeState; the executor reads it (single-loop seeds the saved transcript
// as history; multi-phase skips completed phases and seeds the active phase).
// WithResumeState; the executor reads it:
// - single-loop: History seeds the saved transcript (the run continues).
// - multi-phase: CompletedPhases are skipped; the interrupted phase re-runs
Outdated
Review

🟠 ActivePhase field plumbed through 3 structs but never read/set; doc comments claim active-phase seeding that isn't implemented

correctness, maintainability · flagged by 5 models

  • ActivePhase is dead weight, and its docs overpromise — run/checkpoint.go:25, run/ports.go:210, checkpoint/checkpoint.go:41. The field is plumbed through three structs (ResumeState, RunCheckpointState, RunCheckpoint) but the kernel never reads it and never writes a non-empty value to it. Every write site is either the literal ActivePhase: "" (run/phases.go:186) or a pass-through copy (checkpoint/handle.go:61), and there is no read site anywhere — the single-loop Save (…

🪰 Gadfly · advisory

🟠 **ActivePhase field plumbed through 3 structs but never read/set; doc comments claim active-phase seeding that isn't implemented** _correctness, maintainability · flagged by 5 models_ - **`ActivePhase` is dead weight, and its docs overpromise — `run/checkpoint.go:25`, `run/ports.go:210`, `checkpoint/checkpoint.go:41`.** The field is plumbed through three structs (`ResumeState`, `RunCheckpointState`, `RunCheckpoint`) but the kernel never reads it and never writes a non-empty value to it. Every write site is either the literal `ActivePhase: ""` (`run/phases.go:186`) or a pass-through copy (`checkpoint/handle.go:61`), and there is no read site anywhere — the single-loop `Save` (… <sub>🪰 Gadfly · advisory</sub>
// from its start (boundary-granular — there is no mid-phase transcript
// resume, so History is unused for multi-phase runs).
type ResumeState struct {
History []llm.Message // single-loop transcript OR active-phase transcript
History []llm.Message // single-loop transcript (unused for multi-phase)
CompletedPhases []PhaseOutput // multi-phase: outputs of finished phases, in order
ActivePhase string // multi-phase: the phase that was in flight
}
type resumeStateKey struct{}
3
@@ -79,15 +82,21 @@ func classifyCheckpointOutcome(runErr, cause error) checkpointOutcome {
// finalizeCheckpoint applies the outcome to the per-run checkpointer (nil-safe).
// Runs on a detached context so a cancelled run still records its terminal state.
// Complete/Fail errors are best-effort but logged (a stale record would only
// cause a wasteful boot-recovery retry, not data loss).
Outdated
Review

🔴 finalizeCheckpoint silently ignores Complete/Fail errors leaving stale recovery records

error-handling, security · flagged by 4 models

  • run/checkpoint.go:88-90finalizeCheckpoint swallows errors from cp.Complete and cp.Fail. If the store fails to delete a finished run's checkpoint (network blip, timeout), the record survives and the next boot will incorrectly attempt to resume a run that already terminated. This creates false recovery candidates.

🪰 Gadfly · advisory

🔴 **finalizeCheckpoint silently ignores Complete/Fail errors leaving stale recovery records** _error-handling, security · flagged by 4 models_ - **`run/checkpoint.go:88-90`** — `finalizeCheckpoint` swallows errors from `cp.Complete` and `cp.Fail`. If the store fails to delete a finished run's checkpoint (network blip, timeout), the record survives and the next boot will incorrectly attempt to resume a run that already terminated. This creates false recovery candidates. <sub>🪰 Gadfly · advisory</sub>
func finalizeCheckpoint(ctx context.Context, cp Checkpointer, runErr error, cause error) {
if cp == nil {
return
}
switch classifyCheckpointOutcome(runErr, cause) {
case checkpointComplete:
_ = cp.Complete(detach(ctx))
if err := cp.Complete(detach(ctx)); err != nil {
slog.Warn("run: checkpoint Complete failed", "error", err)
}
case checkpointFail:
_ = cp.Fail(detach(ctx), runErr)
if err := cp.Fail(detach(ctx), runErr); err != nil {
slog.Warn("run: checkpoint Fail failed", "error", err)
}
case checkpointLeaveRunning:
// Interrupted by shutdown: leave the record for boot recovery.
}
+41 -12
View File
@@ -4,6 +4,7 @@ import (
"context"
"errors"
"fmt"
"log/slog"
"time"
"gitea.stevedudenhoeffer.com/steve/majordomo/agent"
@@ -113,13 +114,26 @@ type Result struct {
func (e *Executor) Run(ctx context.Context, ra RunnableAgent, inv tool.Invocation, input string) (res Result) {
started := time.Now()
res = Result{RunID: inv.RunID}
// ckpt is the per-run durable checkpointer (resolved below; nil = non-durable).
// checkpointCause yields the run context's cancellation cause once the run
// context exists; nil before then (an early build-error return).
var ckpt Checkpointer
var checkpointCause func() error
// Enforce the no-panic contract: a panic anywhere in the run (incl. a host
// Critic/Audit/Palette callback on the main goroutine) becomes Result.Err
// rather than unwinding into the caller.
// rather than unwinding into the caller. This defer ALSO finalizes the
// checkpoint on EVERY exit path — panic, an early build-error return (before
// the run loop), or normal completion — so a recovered run's durable record is
// never left dangling (which would loop boot-recovery on a persistent error).
defer func() {
if r := recover(); r != nil {
res.Err = fmt.Errorf("run.Executor: recovered panic: %v", r)
}
var cause error
if checkpointCause != nil {
cause = checkpointCause()
}
finalizeCheckpoint(ctx, ckpt, res.Err, cause)
}()
tier := ra.ModelTier
2
@@ -188,9 +202,15 @@ func (e *Executor) Run(ctx context.Context, ra RunnableAgent, inv tool.Invocatio
// Checkpointer via the factory (which decides durability — nil = non-durable).
// nil-safe throughout.
resume := resumeStateFromContext(ctx)
ckpt := existingCheckpointerFromContext(ctx)
ckpt = existingCheckpointerFromContext(ctx)
if ckpt == nil && e.cfg.Ports.Checkpointer != nil {
if c, cerr := e.cfg.Ports.Checkpointer.Begin(ctx, info); cerr == nil {
c, cerr := e.cfg.Ports.Checkpointer.Begin(ctx, info)
if cerr != nil {
// Degrade to non-durable (the documented contract) but log it — a
// failing checkpoint store must not fail the run, yet shouldn't be silent.
slog.Warn("run: checkpointer Begin failed; running non-durable",
"run_id", inv.RunID, "error", cerr)
} else {
ckpt = c
}
}
@@ -262,6 +282,9 @@ func (e *Executor) Run(ctx context.Context, ra RunnableAgent, inv tool.Invocatio
defer cancelCause(nil)
runCtx, mergeCancel := MergeCancellation(runCtx, ctx)
defer mergeCancel()
// The finalize defer (top of Run) now has a run context to read the
// cancellation cause from (shutdown vs critic-kill vs deadline vs cancel).
checkpointCause = func() error { return context.Cause(runCtx) }
// Critic (optional): monitors the run for a stall, can nudge/extend/kill via
// its host Escalator. Its hard deadline is bound to runCtx (cancel on pass).
1
@@ -339,6 +362,8 @@ func (e *Executor) Run(ctx context.Context, ra RunnableAgent, inv tool.Invocatio
// the critic's nudges before each step.
steer := func() []llm.Message { return append(mailbox.drain(), critic.drainSteer()...) }
resuming := resume != nil && len(resume.History) > 0
var runRes *agent.Result
Outdated
Review

🔴 Per-step checkpoint Save error silently swallowed

correctness, error-handling, performance, security · flagged by 6 models

  • run/executor.go:367 — Per-step checkpoint Save error silently swallowed: The per-step save in the single-loop path discards the error with _ =. Same impact as above — checkpoint persistence failures go unnoticed, risking stale recovery state. Verified by reading run/executor.go:359-368. Suggested fix: Log the error; at minimum ensure the failure is recorded in the audit trail.

🪰 Gadfly · advisory

🔴 **Per-step checkpoint Save error silently swallowed** _correctness, error-handling, performance, security · flagged by 6 models_ - **`run/executor.go:367` — Per-step checkpoint `Save` error silently swallowed**: The per-step save in the single-loop path discards the error with `_ =`. Same impact as above — checkpoint persistence failures go unnoticed, risking stale recovery state. **Verified** by reading `run/executor.go:359-368`. *Suggested fix*: Log the error; at minimum ensure the failure is recorded in the audit trail. <sub>🪰 Gadfly · advisory</sub>
var runErr error
if len(ra.Phases) == 0 {
@@ -347,14 +372,19 @@ func (e *Executor) Run(ctx context.Context, ra RunnableAgent, inv tool.Invocatio
// healthy-but-long run's budget mid-flight; falls back to maxIter).
//
// Checkpointing: wrap the step observer to accumulate the running transcript
// and Save it each step (the host throttles). A recovered run seeds the saved
// transcript as history and continues with no new input. acc starts from the
// resume history (or the opening user message) and grows as steps complete.
// and Save it each step. Save is called every step; THROTTLING is the
// Checkpointer's responsibility (the battery + mort's durable-job adapter
// both throttle + size-cap), so the kernel doesn't gate the hot path. The
// accumulated transcript is the pre-compaction one (the observer sees raw
// step responses, not the loop's compacted history) — a host that caps size
// bounds it. A recovered run seeds the saved transcript and continues.
obs := stepObserver
if ckpt != nil {
acc := []llm.Message{multimodalUserMessage(input, inv.Images)}
if resume != nil && len(resume.History) > 0 {
var acc []llm.Message
if resuming {
acc = append([]llm.Message(nil), resume.History...)
} else {
acc = []llm.Message{multimodalUserMessage(input, inv.Images)}
}
obs = func(s agent.Step) {
stepObserver(s)
@@ -373,7 +403,7 @@ func (e *Executor) Run(ctx context.Context, ra RunnableAgent, inv tool.Invocatio
agent.WithStepObserver(obs),
}, sharedOpts...)
Outdated
Review

🟠 finalizeCheckpoint not deferred: a panic (or early setup-error return) bypasses it, leaving an orphaned recoverable checkpoint that should be Fail'd → boot recovery loop

error-handling · flagged by 1 model

  • run/executor.go:404finalizeCheckpoint is a plain statement, not deferred, so panics and early-error returns bypass it; a panicked run leaves an orphaned recoverable checkpoint. The checkpointer is begun at :193, the wrapped step observer Saves a snapshot each completed step (:367), but finalizeCheckpoint(...) only runs at :404 after the dispatch. The top-level recover() defer (:119-123) converts a panic anywhere in the loop into res.Err and returns without finali…

🪰 Gadfly · advisory

🟠 **finalizeCheckpoint not deferred: a panic (or early setup-error return) bypasses it, leaving an orphaned recoverable checkpoint that should be Fail'd → boot recovery loop** _error-handling · flagged by 1 model_ - **`run/executor.go:404` — `finalizeCheckpoint` is a plain statement, not deferred, so panics and early-error returns bypass it; a panicked run leaves an orphaned recoverable checkpoint.** The checkpointer is begun at `:193`, the wrapped step observer `Save`s a snapshot each completed step (`:367`), but `finalizeCheckpoint(...)` only runs at `:404` after the dispatch. The top-level `recover()` defer (`:119-123`) converts a panic anywhere in the loop into `res.Err` and returns **without** finali… <sub>🪰 Gadfly · advisory</sub>
ag := agent.New(model, e.systemPrompt(ra), opts...)
if resume != nil && len(resume.History) > 0 {
if resuming {
// Resume: seed the saved transcript and continue (no new input — the
// completed tool calls in the transcript are NOT re-run).
runRes, runErr = ag.Run(runCtx, "", agent.WithSteer(steer), agent.WithHistory(resume.History))
@@ -399,9 +429,8 @@ func (e *Executor) Run(ctx context.Context, ra RunnableAgent, inv tool.Invocatio
}, input, inv.Images)
}
// Finalize durable recovery: clear the checkpoint on success/terminal failure,
// or leave it for boot recovery when the run was interrupted by shutdown.
finalizeCheckpoint(ctx, ckpt, runErr, context.Cause(runCtx))
// Durable-recovery finalize (Complete/Fail/leave-running) happens in the
// top-of-Run defer so it covers panics + early build-error returns too.
status := statusFor(runCtx, runErr)
if runRes != nil {
+15 -11
View File
@@ -53,9 +53,10 @@ import (
// phaseDeps carries the per-run state the phase runner shares with Run: the base
// model, the full decorated toolbox (filtered per phase), the base step cap, the
// shared agent options (tool-error limits + step observer + compactor), the
// shared step observer (also fed by IsRunFunc bare calls), the critic/session
// steer, and the audit recorder (phase events).
// shared agent options (tool-error limits + compactor — the step observer is
// added per phase, NOT in sharedOpts, so checkpointing can vary per path), the
// shared step observer (wired into each phase's loop AND invoked for IsRunFunc
// bare calls), the critic/session steer, and the audit recorder (phase events).
type phaseDeps struct {
baseModel llm.Model
baseToolbox *llm.Toolbox
2
@@ -85,12 +86,18 @@ func (e *Executor) runPhases(runCtx context.Context, ra RunnableAgent, deps phas
var lastOutput string
var totalUsage llm.Usage
// Resume: pre-populate from the saved checkpoint so already-finished phases are
// skipped. The interrupted (active) phase is NOT pre-populated, so it re-runs
// from its start (boundary-granular recovery).
// resumeSkip is the set of phases already finished on a RECOVERED run — kept
// SEPARATE from the live `outputs` map (which fills as phases run this time) so
// the skip guard only skips RESUME-completed phases, never a fresh run's own
// phases. (Reusing `outputs` would make a second phase with a duplicate name
// skip itself.) Pre-populate outputs + completed so a resumed run threads the
// saved outputs into later phases. The interrupted (active) phase is NOT
// pre-populated, so it re-runs from its start (boundary-granular recovery).
resumeSkip := map[string]bool{}
if deps.resume != nil {
for _, pc := range deps.resume.CompletedPhases {
outputs[pc.Name] = pc.Output
resumeSkip[pc.Name] = true
completed = append(completed, pc)
lastOutput = pc.Output
}
1
@@ -109,10 +116,8 @@ func (e *Executor) runPhases(runCtx context.Context, ra RunnableAgent, deps phas
}
for i, phase := range ra.Phases {
// Skip phases already completed on a resumed run (key presence, not output
// emptiness — a legitimately-empty phase output still counts as done).
if _, done := outputs[phase.Name]; done {
lastOutput = outputs[phase.Name]
// Skip phases already completed on a resumed run.
if resumeSkip[phase.Name] {
continue
}
// A killed/timed-out/cancelled run must not start its next phase.
1
@@ -183,7 +188,6 @@ func (e *Executor) runPhases(runCtx context.Context, ra RunnableAgent, deps phas
if deps.checkpointer != nil {
_ = deps.checkpointer.Save(runCtx, RunCheckpointState{
CompletedPhases: append([]PhaseOutput(nil), completed...),
ActivePhase: "",
})
}
}
+20
View File
@@ -130,6 +130,26 @@ func TestPhases_OptionalDoesNotSwallowCancellation(t *testing.T) {
}
}
// TestPhases_DuplicateNamesBothRun: a fresh (non-resume) run with two phases
// sharing a name must run BOTH — the resume-skip guard keys off a separate
// resume set, not the live outputs map (which fills as phases run), so a phase
// never skips a same-named sibling on a fresh run.
func TestPhases_DuplicateNamesBothRun(t *testing.T) {
models, fp := phaseProvider(t, fake.Reply("first"), fake.Reply("second"))
ex := New(Config{Registry: tool.NewRegistry(), Models: models})
ra := RunnableAgent{
Name: "p", ModelTier: "test-model",
Phases: []Phase{{Name: "x", SystemPrompt: "P1"}, {Name: "x", SystemPrompt: "P2"}},
}
res := ex.Run(context.Background(), ra, tool.Invocation{RunID: "r"}, "Q")
if res.Err != nil {
t.Fatalf("run error: %v", res.Err)
}
if n := len(fp.Calls()); n != 2 {
t.Fatalf("both same-named phases must run on a fresh run; got %d model calls", n)
}
}
// TestPhases_HardErrorAborts: a NON-optional phase that hits a hard error (not a
// budget/step exhaustion) aborts the pipeline; later phases do not run.
func TestPhases_HardErrorAborts(t *testing.T) {
+5 -5
View File
@@ -199,15 +199,15 @@ type Checkpointer interface {
// RunCheckpointState is the resumable snapshot a Checkpointer persists.
type RunCheckpointState struct {
// Messages is the running transcript (single-loop run) OR the active phase's
// transcript (multi-phase run). May be nil.
// Messages is the running transcript of a SINGLE-LOOP run (grows each step;
Outdated
Review

🟡 Documentation claims multi-phase transcript resume not implemented

correctness · flagged by 1 model

  1. run/ports.go:202-203 — Documentation claims unsupported functionality - The comment states Messages is "the running transcript (single-loop run) OR the active phase's transcript (multi-phase run)". However, the multi-phase checkpoint save at run/phases.go:184-187 never sets Messages — it only saves CompletedPhases and ActivePhase. This means mid-phase transcript recovery for multi-phase runs is NOT supported (only boundary-granular). - Impact: Documentation promises fu…

🪰 Gadfly · advisory

🟡 **Documentation claims multi-phase transcript resume not implemented** _correctness · flagged by 1 model_ 2. **`run/ports.go:202-203` — Documentation claims unsupported functionality** - The comment states `Messages` is "the running transcript (single-loop run) OR the active phase's transcript (multi-phase run)". However, the multi-phase checkpoint save at `run/phases.go:184-187` **never sets `Messages`** — it only saves `CompletedPhases` and `ActivePhase`. This means mid-phase transcript recovery for multi-phase runs is NOT supported (only boundary-granular). - **Impact**: Documentation promises fu… <sub>🪰 Gadfly · advisory</sub>
// resumed via WithHistory). nil for multi-phase runs — phase recovery is
// boundary-granular (see CompletedPhases), not mid-phase transcript.
Messages []llm.Message
Iteration int
// CompletedPhases is set only for multi-phase runs: the outputs of phases
// already finished, in phase order. nil for single-loop runs.
// already finished, in phase order, so a resumed run skips them and re-runs
// the interrupted phase from its start. nil for single-loop runs.
CompletedPhases []PhaseOutput
Review

🟡 RunCheckpointState.ActivePhase is always empty string — field exists but is never populated with meaningful data

maintainability · flagged by 1 model

  1. run/ports.go:210 / run/phases.go:186RunCheckpointState.ActivePhase is defined and persisted, but the only write in the entire codebase is ActivePhase: "" at run/phases.go:186. It is never set to a non-empty value.

🪰 Gadfly · advisory

🟡 **RunCheckpointState.ActivePhase is always empty string — field exists but is never populated with meaningful data** _maintainability · flagged by 1 model_ 2. **`run/ports.go:210` / `run/phases.go:186`** — `RunCheckpointState.ActivePhase` is defined and persisted, but the only write in the entire codebase is `ActivePhase: ""` at `run/phases.go:186`. It is never set to a non-empty value. <sub>🪰 Gadfly · advisory</sub>
// ActivePhase is the name of the in-progress phase (multi-phase only).
ActivePhase string
}
// PhaseOutput is one completed pipeline phase's name and output text, recorded in