Heartbeat firing twice with different models
So, here is something new I noticed today after trying to switch my heartbeat model to ollama. From my log files, it looks like the heartbeat is firing twice - or trying to, once with ollama and once with my fallback openai model. Same runId...same sessionId. Asked my claw to self-diagnose and it ultimately suggested we kill the fallback model. Which...ok, but...
From logs:
11:17:02 PM
debug
agent/embedded
embedded run start: runId=ba4314c4-39fc-4354-b0f1-7f99945b7f01 sessionId=cf212a91-1e1d-4ef4-9338-4a399a380d27 provider=ollama model=llama3.2:3b thinking=off messageChannel=heartbeat
11:27:02 PM
debug
agent/embedded
embedded run start: runId=ba4314c4-39fc-4354-b0f1-7f99945b7f01 sessionId=cf212a91-1e1d-4ef4-9338-4a399a380d27 provider=openai model=gpt-5.3-codex thinking=off messageChannel=heartbeat
There are approximatley 12 hours of entries like this. Each one exactly 10 minutes apart.
From config:
"agents": {
"defaults": {
"model": {
"primary": "anthropic/claude-haiku-4-5",
"fallbacks": [
"anthropic/claude-opus-4-6",
"openai/gpt-5.3-codex"
]
},
"models": {
"openai/gpt-5.3-codex": {
"alias": "gpt"
},
"anthropic/claude-opus-4-6": {
"alias": "opus"
},
"anthropic/claude-haiku-4-5": {
"alias": "haiku"
},
"ollama/llama3.2:3b": {
"alias": "llama"
}
},
"workspace": "~/.openclaw/workspace",
"compaction": {
"mode": "safeguard"
},
"heartbeat": {
"every": "30m",
"model": "ollama/llama3.2:3b",
"includeReasoning": false,
"target": "last",
"ackMaxChars": 300
},
"maxConcurrent": 4,
"subagents": {
"maxConcurrent": 8
}
}
},
Has anyone else seen this?
From logs:
11:17:02 PM
debug
agent/embedded
embedded run start: runId=ba4314c4-39fc-4354-b0f1-7f99945b7f01 sessionId=cf212a91-1e1d-4ef4-9338-4a399a380d27 provider=ollama model=llama3.2:3b thinking=off messageChannel=heartbeat
11:27:02 PM
debug
agent/embedded
embedded run start: runId=ba4314c4-39fc-4354-b0f1-7f99945b7f01 sessionId=cf212a91-1e1d-4ef4-9338-4a399a380d27 provider=openai model=gpt-5.3-codex thinking=off messageChannel=heartbeat
There are approximatley 12 hours of entries like this. Each one exactly 10 minutes apart.
From config:
"agents": {
"defaults": {
"model": {
"primary": "anthropic/claude-haiku-4-5",
"fallbacks": [
"anthropic/claude-opus-4-6",
"openai/gpt-5.3-codex"
]
},
"models": {
"openai/gpt-5.3-codex": {
"alias": "gpt"
},
"anthropic/claude-opus-4-6": {
"alias": "opus"
},
"anthropic/claude-haiku-4-5": {
"alias": "haiku"
},
"ollama/llama3.2:3b": {
"alias": "llama"
}
},
"workspace": "~/.openclaw/workspace",
"compaction": {
"mode": "safeguard"
},
"heartbeat": {
"every": "30m",
"model": "ollama/llama3.2:3b",
"includeReasoning": false,
"target": "last",
"ackMaxChars": 300
},
"maxConcurrent": 4,
"subagents": {
"maxConcurrent": 8
}
}
},
Has anyone else seen this?
The AI that actually does things. Emails, calendar, home automation - all from your favorite chat app. New shell, ready to help. The lobster way.
134,426Members
Resources
Similar Threads
Was this page helpful?
