Nikolay Tatarinov
6a27fecb13
All checks were successful
CI / ci (push) Successful in 14s
- Introduced a new `stats.py` module to encapsulate dashboard statistics building and cache key constants. - Refactored `views.py` to utilize the new `build_stats` function for constructing metrics context, improving code organization and readability. - Updated Prometheus query handling to streamline metrics fetching with a new `fetch_dashboard_metrics` function. - Enhanced test cases to reflect changes in metrics fetching and context building, ensuring accurate functionality. - Added new HTML templates for displaying detailed resource allocation and flavor statistics on the dashboard.
87 lines
3.4 KiB
Python
87 lines
3.4 KiB
Python
from concurrent.futures import ThreadPoolExecutor, as_completed
|
|
|
|
import requests
|
|
from watcher_visio.settings import PROMETHEUS_URL
|
|
|
|
# Timeout for lightweight health check (seconds)
|
|
CHECK_TIMEOUT = 5
|
|
|
|
# Dashboard Prometheus queries (query_key -> query string), run in parallel
|
|
DASHBOARD_QUERIES = {
|
|
"hosts_total": "count(node_exporter_build_info{job='node_exporter_compute'})",
|
|
"pcpu_total": (
|
|
"sum(count(node_cpu_seconds_total{job='node_exporter_compute', mode='idle'}) "
|
|
"without (cpu,mode))"
|
|
),
|
|
"pcpu_usage": "sum(node_load5{job='node_exporter_compute'})",
|
|
"vcpu_allocated": "sum(libvirt_domain_info_virtual_cpus)",
|
|
"vcpu_overcommit_max": (
|
|
"avg(openstack_placement_resource_allocation_ratio{resourcetype='VCPU'})"
|
|
),
|
|
"pram_total": "sum(node_memory_MemTotal_bytes{job='node_exporter_compute'})",
|
|
"pram_usage": "sum(node_memory_Active_bytes{job='node_exporter_compute'})",
|
|
"vram_allocated": "sum(libvirt_domain_info_maximum_memory_bytes)",
|
|
"vram_overcommit_max": (
|
|
"avg(avg_over_time("
|
|
"openstack_placement_resource_allocation_ratio{resourcetype='MEMORY_MB'}[5m]))"
|
|
),
|
|
"vm_count": "sum(libvirt_domain_state_code)",
|
|
"vm_active": "sum(libvirt_domain_state_code{stateDesc='the domain is running'})",
|
|
}
|
|
|
|
# Keys that should be parsed as float (rest as int)
|
|
DASHBOARD_FLOAT_KEYS = frozenset(("pcpu_usage", "vcpu_overcommit_max", "vram_overcommit_max"))
|
|
|
|
|
|
def check_prometheus() -> dict:
|
|
"""
|
|
Lightweight check that Prometheus is reachable.
|
|
Returns {"status": "ok"} or {"status": "error", "message": "..."}.
|
|
"""
|
|
url = f"{PROMETHEUS_URL.rstrip('/')}/api/v1/query"
|
|
try:
|
|
response = requests.get(url, params={"query": "1"}, timeout=CHECK_TIMEOUT)
|
|
response.raise_for_status()
|
|
data = response.json()
|
|
if "data" in data and "result" in data["data"]:
|
|
return {"status": "ok"}
|
|
return {"status": "error", "message": "Invalid response"}
|
|
except requests.RequestException as e:
|
|
return {"status": "error", "message": str(e) or "Connection failed"}
|
|
except (ValueError, KeyError) as e:
|
|
return {"status": "error", "message": str(e) or "Invalid response"}
|
|
|
|
|
|
def query_prometheus(query: str) -> str | list[str]:
|
|
url = f"{PROMETHEUS_URL}/api/v1/query"
|
|
params = {
|
|
"query": query,
|
|
}
|
|
response = requests.get(url=url, params=params, timeout=CHECK_TIMEOUT)
|
|
response.raise_for_status()
|
|
result = response.json()["data"]["result"]
|
|
if len(result) > 1:
|
|
return result
|
|
else:
|
|
return result[0]["value"][1]
|
|
|
|
|
|
def fetch_dashboard_metrics() -> dict:
|
|
"""Run all dashboard Prometheus queries in parallel and return a dict of name -> value."""
|
|
result = {}
|
|
with ThreadPoolExecutor(max_workers=len(DASHBOARD_QUERIES)) as executor:
|
|
future_to_key = {
|
|
executor.submit(query_prometheus, query=q): key for key, q in DASHBOARD_QUERIES.items()
|
|
}
|
|
for future in as_completed(future_to_key):
|
|
key = future_to_key[future]
|
|
try:
|
|
raw = future.result()
|
|
if key in DASHBOARD_FLOAT_KEYS:
|
|
result[key] = float(raw)
|
|
else:
|
|
result[key] = int(raw)
|
|
except (ValueError, TypeError):
|
|
result[key] = float(0) if key in DASHBOARD_FLOAT_KEYS else 0
|
|
return result
|