zzstoatzz.io
prefect server in zig
update roadmap with implemented features
- database backends (sqlite + postgres)
- message broker (memory + redis)
- event_broadcaster service
- postgres connection pooling
- updated priority notes
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
+27
-4
ROADMAP.md
+27
-4
ROADMAP.md
···
106
106
## background services
107
107
108
108
- [x] event_persister (batched event writes, deduplication, retention trimming)
109
+
- [x] event_broadcaster (websocket fan-out to /api/events/out subscribers)
109
110
- [ ] scheduler (create flow runs from deployment schedules)
110
111
- [ ] late_runs (mark runs as late)
111
112
- [ ] foreman (infrastructure management)
···
113
114
- [ ] pause_expirations (expire paused runs)
114
115
- [ ] task_run_recorder (record task run events)
115
116
- [ ] telemetry
117
+
118
+
## message broker
119
+
120
+
- [x] memory broker (growable queue, in-process)
121
+
- [x] redis broker (redis streams: XADD/XREADGROUP/XACK)
122
+
- [x] broker abstraction layer (broker/core.zig)
123
+
- [x] fan-out to consumer groups
116
124
117
125
## database
118
126
···
135
143
- [ ] variable table
136
144
- [ ] log table (currently in-memory only)
137
145
146
+
### database backends
147
+
- [x] sqlite (zqlite)
148
+
- [x] postgres (pg.zig)
149
+
- [x] backend abstraction layer (db/backend.zig)
150
+
138
151
## infrastructure
139
152
140
153
- [x] http server (zap/facil.io)
141
154
- [x] websocket support
142
155
- [x] sqlite database (zqlite)
156
+
- [x] postgres database (pg.zig)
143
157
- [x] structured logging
144
158
- [x] service lifecycle management
159
+
- [x] postgres connection pooling (pg.zig built-in)
160
+
- [ ] sqlite connection pooling
145
161
- [ ] migrations
146
-
- [ ] connection pooling
147
162
- [ ] graceful shutdown
148
163
- [ ] configuration file support
149
164
150
165
## notes
151
166
152
167
priority order for next work:
153
-
1. deployments (needed for scheduled/triggered runs)
154
-
2. work_pools (needed for worker-based execution)
155
-
3. scheduler service (needed for deployment schedules)
168
+
1. variables (simple CRUD, useful standalone, exercises db layer)
169
+
2. deployments (needed for scheduled/triggered runs)
170
+
3. work_pools (needed for worker-based execution)
171
+
4. scheduler service (needed for deployment schedules)
172
+
173
+
### what's working (6.5x faster than python, 3.8x less memory)
174
+
- flow/flow_run/task_run lifecycle
175
+
- blocks (types, schemas, documents)
176
+
- events (ingest via websocket, persist, broadcast)
177
+
- dual database backends (sqlite/postgres)
178
+
- dual message brokers (memory/redis)