Shorten agent health checks on load

This commit is contained in:
James Smith
2026-03-19 08:09:07 +00:00
parent db0a26cd64
commit fb8b6a01e8
2 changed files with 80 additions and 53 deletions

View File

@@ -40,14 +40,16 @@ from utils.trilateration import (
estimate_location_from_observations,
)
logger = logging.getLogger('intercept.controller')
controller_bp = Blueprint('controller', __name__, url_prefix='/controller')
# Multi-agent SSE fanout state (per-client queues).
_agent_stream_subscribers: set[queue.Queue] = set()
_agent_stream_subscribers_lock = threading.Lock()
_AGENT_STREAM_CLIENT_QUEUE_SIZE = 500
logger = logging.getLogger('intercept.controller')
controller_bp = Blueprint('controller', __name__, url_prefix='/controller')
AGENT_HEALTH_TIMEOUT_SECONDS = 2.0
AGENT_STATUS_TIMEOUT_SECONDS = 2.5
# Multi-agent SSE fanout state (per-client queues).
_agent_stream_subscribers: set[queue.Queue] = set()
_agent_stream_subscribers_lock = threading.Lock()
_AGENT_STREAM_CLIENT_QUEUE_SIZE = 500
def _broadcast_agent_data(payload: dict) -> None:
@@ -77,14 +79,18 @@ def get_agents():
agents = list_agents(active_only=active_only)
# Optionally refresh status for each agent
refresh = request.args.get('refresh', 'false').lower() == 'true'
if refresh:
for agent in agents:
try:
client = create_client_from_agent(agent)
agent['healthy'] = client.health_check()
except Exception:
agent['healthy'] = False
refresh = request.args.get('refresh', 'false').lower() == 'true'
if refresh:
for agent in agents:
try:
client = AgentClient(
agent['base_url'],
api_key=agent.get('api_key'),
timeout=AGENT_HEALTH_TIMEOUT_SECONDS,
)
agent['healthy'] = client.health_check()
except Exception:
agent['healthy'] = False
return jsonify({
'status': 'success',
@@ -327,27 +333,36 @@ def check_all_agents_health():
'error': None
}
try:
client = create_client_from_agent(agent)
# Time the health check
start_time = time.time()
is_healthy = client.health_check()
response_time = (time.time() - start_time) * 1000
try:
client = AgentClient(
agent['base_url'],
api_key=agent.get('api_key'),
timeout=AGENT_HEALTH_TIMEOUT_SECONDS,
)
# Time the health check
start_time = time.time()
is_healthy = client.health_check()
response_time = (time.time() - start_time) * 1000
result['healthy'] = is_healthy
result['response_time_ms'] = round(response_time, 1)
if is_healthy:
# Update last_seen in database
update_agent(agent['id'], update_last_seen=True)
# Also fetch running modes
try:
status = client.get_status()
result['running_modes'] = status.get('running_modes', [])
result['running_modes_detail'] = status.get('running_modes_detail', {})
except Exception:
# Update last_seen in database
update_agent(agent['id'], update_last_seen=True)
# Also fetch running modes
try:
status_client = AgentClient(
agent['base_url'],
api_key=agent.get('api_key'),
timeout=AGENT_STATUS_TIMEOUT_SECONDS,
)
status = status_client.get_status()
result['running_modes'] = status.get('running_modes', [])
result['running_modes_detail'] = status.get('running_modes_detail', {})
except Exception:
pass # Status fetch is optional
except AgentConnectionError as e:

View File

@@ -10,10 +10,11 @@ let currentAgent = 'local';
let agentEventSource = null;
let multiAgentMode = false; // Show combined results from all agents
let multiAgentPollInterval = null;
let agentRunningModes = []; // Track agent's running modes for conflict detection
let agentRunningModesDetail = {}; // Track device info per mode (for multi-SDR agents)
let healthCheckInterval = null; // Health monitoring interval
let agentHealthStatus = {}; // Cache of health status per agent ID
let agentRunningModes = []; // Track agent's running modes for conflict detection
let agentRunningModesDetail = {}; // Track device info per mode (for multi-SDR agents)
let healthCheckInterval = null; // Health monitoring interval
let agentHealthStatus = {}; // Cache of health status per agent ID
let healthCheckKickoffTimer = null;
// ============== AGENT HEALTH MONITORING ==============
@@ -21,27 +22,38 @@ let agentHealthStatus = {}; // Cache of health status per agent ID
* Start periodic health monitoring for all agents.
* Runs every 30 seconds to check agent health status.
*/
function startHealthMonitoring() {
// Don't start if already running
if (healthCheckInterval) return;
// Initial check
checkAllAgentsHealth();
// Start periodic checks every 30 seconds
healthCheckInterval = setInterval(checkAllAgentsHealth, 30000);
console.log('[AgentManager] Health monitoring started (30s interval)');
}
function startHealthMonitoring() {
// Don't start if already running
if (healthCheckInterval) return;
// Defer the first probe so heavy dashboards can finish initial render
// before we start contacting remote agents.
if (healthCheckKickoffTimer) {
clearTimeout(healthCheckKickoffTimer);
}
healthCheckKickoffTimer = setTimeout(() => {
healthCheckKickoffTimer = null;
checkAllAgentsHealth();
}, 5000);
// Start periodic checks every 30 seconds
healthCheckInterval = setInterval(checkAllAgentsHealth, 30000);
console.log('[AgentManager] Health monitoring started (30s interval)');
}
/**
* Stop health monitoring.
*/
function stopHealthMonitoring() {
if (healthCheckInterval) {
clearInterval(healthCheckInterval);
healthCheckInterval = null;
console.log('[AgentManager] Health monitoring stopped');
}
function stopHealthMonitoring() {
if (healthCheckKickoffTimer) {
clearTimeout(healthCheckKickoffTimer);
healthCheckKickoffTimer = null;
}
if (healthCheckInterval) {
clearInterval(healthCheckInterval);
healthCheckInterval = null;
console.log('[AgentManager] Health monitoring stopped');
}
}
/**