Logflare | Cloudflare, Vercel & Elixir Logging

~/ endpoints / functions.invocation-stats /edit

docs

Endpoint Name #

This endpoint will can be queried at:
http://localhost:4000/api/endpoints/query/YourApp.EndpointName

Query #

This endpoint will execute this query on each GET request. Declare parameters using @parameter_name and pass it to the endpoint via query parameters ?parameter_name=some-value

with arr as (
  SELECT  (
      CASE
        WHEN @interval = '5min' THEN  `supabase-analytics-eu.logflare.generate_timestamp_trunc_array`(5, "minute_second")
        WHEN @interval = '15min' THEN  `supabase-analytics-eu.logflare.generate_timestamp_trunc_array`(15, "minute_minute")
        WHEN @interval = '1hr' THEN `supabase-analytics-eu.logflare.generate_timestamp_trunc_array`(1, "hour_minute")
        WHEN @interval = '1day' THEN `supabase-analytics-eu.logflare.generate_timestamp_trunc_array`(1, "day_hour")
        WHEN @interval = '7day' THEN `supabase-analytics-eu.logflare.generate_timestamp_trunc_array`(7, "day_day")
        ELSE `supabase-analytics-eu.logflare.generate_timestamp_trunc_array`(1, "day_hour")
      END
  ) as d
),
dates as (
  select d from arr, unnest(`arr`.d) as d
),
agg as (
SELECT 
  dates.d as timestamp,
  count(id) as count,
  avg(m.execution_time_ms) as avg_execution_time,
  max(m.execution_time_ms) as max_execution_time,
  min(m.execution_time_ms) as min_execution_time,
  COUNTIF(r.status_code >= 400) as error_count,
  COUNTIF(r.status_code >= 100 AND r.status_code < 200 ) as one_xx_status_code_count,
  COUNTIF(r.status_code >= 200 AND r.status_code < 300 ) as two_xx_status_code_count,
  COUNTIF(r.status_code >= 300 AND r.status_code < 400 ) as three_xx_status_code_count,
  COUNTIF(r.status_code >= 400 AND r.status_code < 500 ) as four_xx_status_code_count,
  COUNTIF(r.status_code >= 500 AND r.status_code < 600 ) as five_xx_status_code_count,
  APPROX_QUANTILES(m.execution_time_ms, 100) as p95_array,
  APPROX_QUANTILES(m.execution_time_ms, 100) as p99_array,
FROM
  arr, dates left join `deno-relay-logs` on dates.d = (
        CASE
            WHEN @interval = '5min' THEN timestamp_trunc(timestamp, second)
            WHEN @interval = '15min' THEN timestamp_trunc(timestamp, minute)
            WHEN @interval = '1hr' THEN timestamp_trunc(timestamp, minute)
            WHEN @interval = '1day' THEN timestamp_trunc(timestamp, hour)
            WHEN @interval = '7day' THEN timestamp_trunc(timestamp, day)
        ELSE timestamp_trunc(timestamp, hour)
      END)
CROSS JOIN UNNEST(metadata) as m
CROSS JOIN UNNEST(m.response) as r
WHERE
  m.project_ref = @project 
  AND CASE WHEN COALESCE(@function_id, '') = '' THEN  TRUE ELSE  m.function_id = @function_id END
  and timestamp >= arr.d[offset(0)]
GROUP BY
  timestamp
ORDER BY
  timestamp ASC
)

select
  timestamp,
  min_execution_time,
  max_execution_time,
  avg_execution_time,
  p95_array[offset(95)] as p95_execution_time,
  p99_array[offset(99)] as p99_execution_time,
  count,
  error_count
from
  agg

Max limit

rows

This option determines the maximium result size returned from BigQuery.

Enable authentication Endpoints with authentication enabled will require an access token to be provided on each request.

Enable query sandboxing Endpoints with sandboxing enabled will accept an ?sql= query parameter that will be inserted as the outer query of the defined CTE. This restricts users of the endpoint to only query within the defined CTE result sets.

Caching #

Cache TTL

seconds

Endpoint results are cached on a per-request basis. Set Cache TTL to 0 to disable caching completely.

Proactive Re-querying

seconds

Cached results will be refreshed at this interval. Ignored if caching is disabled.