Read GKE logs in ICANN reports (#2673)

GKE logs are routed to a different dataset and the table is different.
The structs to look for are also different (jsonPayload vs textPayload
or protoPayload).

TESTED=Ran the resulting query in crash.
This commit is contained in:
Lai Jiang
2025-02-12 20:41:44 +00:00
committed by GitHub
parent c918258fb1
commit e78de98060
16 changed files with 164 additions and 85 deletions
@@ -19,7 +19,7 @@
-- monthly App Engine logs, searching for all create commands and associating
-- them with their corresponding registrars.
-- Example log generated by FlowReporter in App Engine logs:
-- Example log generated by FlowReporter in App Engine and GKE logs:
--google.registry.flows.FlowReporter
-- recordToLogs: FLOW-LOG-SIGNATURE-METADATA:
--{"serverTrid":"oNwL2J2eRya7bh7c9oHIzg==-2360a","clientId":"ipmirror"
@@ -38,30 +38,42 @@ FROM (
JSON_EXTRACT_SCALAR(json, '$.tld') AS tld,
JSON_EXTRACT_SCALAR(json, '$.clientId') AS clientId,
COUNT(json) AS count
FROM (
FROM ((
-- Extract JSON metadata package from monthly logs
SELECT
REGEXP_EXTRACT(textPayload, r'FLOW-LOG-SIGNATURE-METADATA: (.*)\n?$')
AS json
FROM (
REGEXP_EXTRACT(textPayload, r'FLOW-LOG-SIGNATURE-METADATA: (.*)\n?$') AS json
FROM
`domain-registry-alpha.appengine_logs._var_log_app_*`
WHERE
_TABLE_SUFFIX BETWEEN '20170901'
AND '20170930'
AND STARTS_WITH(textPayload, "FLOW-LOG-SIGNATURE-METADATA")
-- Look for domain creates
AND REGEXP_CONTAINS(textPayload, r'"commandType":"create","resourceType":"domain"')
-- Filter prober data
AND NOT REGEXP_CONTAINS(textPayload, r'"prober-[a-z]{2}-((any)|(canary))"'))
UNION ALL (
-- Extract JSON metadata package from monthly logs
SELECT
textPayload
REGEXP_EXTRACT(jsonPayload.message, r'FLOW-LOG-SIGNATURE-METADATA: (.*)\n?$') AS json
FROM
`domain-registry-alpha.appengine_logs._var_log_app_*`
WHERE _TABLE_SUFFIX
BETWEEN '20170901'
AND '20170930')
WHERE STARTS_WITH(textPayload, "FLOW-LOG-SIGNATURE-METADATA")
-- Look for domain creates
AND REGEXP_CONTAINS(
textPayload, r'"commandType":"create","resourceType":"domain"')
-- Filter prober data
AND NOT REGEXP_CONTAINS(
textPayload, r'"prober-[a-z]{2}-((any)|(canary))"') )
GROUP BY tld, clientId ) AS logs_table
`domain-registry-alpha.gke_logs.stderr_*`
WHERE
_TABLE_SUFFIX BETWEEN '20170901'
AND '20170930'
AND STARTS_WITH(jsonPayload.message, "FLOW-LOG-SIGNATURE-METADATA")
-- Look for domain creates
AND REGEXP_CONTAINS(jsonPayload.message, r'"commandType":"create","resourceType":"domain"')
-- Filter prober data
AND NOT REGEXP_CONTAINS(jsonPayload.message, r'"prober-[a-z]{2}-((any)|(canary))"')))
GROUP BY
tld,
clientId ) AS logs_table
JOIN
EXTERNAL_QUERY("projects/domain-registry-alpha/locations/us/connections/domain-registry-alpha-sql",
'''SELECT registrar_id, registrar_name FROM "Registrar";''') AS registrar_table
ON
logs_table.clientId = registrar_table.registrar_id
ORDER BY tld, registrar_name
'''SELECT registrar_id, registrar_name FROM "Registrar";''') AS registrar_table
ON
logs_table.clientId = registrar_table.registrar_id
ORDER BY
tld,
registrar_name
@@ -35,14 +35,30 @@ FROM (
JSON_EXTRACT_SCALAR(json,
'$.icannActivityReportField') AS activityReportField
FROM (
-- For reasons that I don't understand, if I directly select the three columns
-- from the union, BigQuery complains about column number mismatch, so I have to
-- make a temporary union table and select on it.
SELECT
-- Extract the logged JSON payload.
REGEXP_EXTRACT(textPayload, r'FLOW-LOG-SIGNATURE-METADATA: (.*)\n?$')
AS json
FROM `domain-registry-alpha.appengine_logs._var_log_app_*`
WHERE
STARTS_WITH(textPayload, "FLOW-LOG-SIGNATURE-METADATA")
AND _TABLE_SUFFIX BETWEEN '20170901' AND '20170930')) AS regexes
*
FROM (
SELECT
-- Extract the logged JSON payload.
REGEXP_EXTRACT(textPayload, r'FLOW-LOG-SIGNATURE-METADATA: (.*)\n?$')
AS json
FROM `domain-registry-alpha.appengine_logs._var_log_app_*`
WHERE
STARTS_WITH(textPayload, "FLOW-LOG-SIGNATURE-METADATA")
AND _TABLE_SUFFIX BETWEEN '20170901' AND '20170930')
UNION ALL (
SELECT
-- Extract the logged JSON payload.
REGEXP_EXTRACT(jsonPayload.message, r'FLOW-LOG-SIGNATURE-METADATA: (.*)\n?$')
AS json
FROM `domain-registry-alpha.gke_logs.stderr_*`
WHERE
STARTS_WITH(jsonPayload.message, "FLOW-LOG-SIGNATURE-METADATA")
AND _TABLE_SUFFIX BETWEEN '20170901' AND '20170930')
)) AS regexes
JOIN
-- Unnest the JSON-parsed tlds.
UNNEST(regexes.tlds) AS tld
@@ -13,13 +13,23 @@
-- See the License for the specific language governing permissions and
-- limitations under the License.
-- Query to fetch AppEngine request logs for the report month.
-- Query to fetch AppEngine and GKE request logs for the report month.
-- START_OF_MONTH and END_OF_MONTH should be in YYYYMM01 format.
SELECT
protoPayload.resource AS requestPath,
FROM
`domain-registry-alpha.appengine_logs.appengine_googleapis_com_request_log_*`
WHERE
_TABLE_SUFFIX BETWEEN '20170901' AND '20170930'
*
FROM (
SELECT
jsonPayload.httrequest.requesturl AS requestPath
FROM
`domain-registry-alpha.gke_logs.stderr_*`
WHERE
_TABLE_SUFFIX BETWEEN '20170901' AND '20170930')
UNION ALL (
SELECT
protoPayload.resource AS requestPath
FROM
`domain-registry-alpha.appengine_logs.appengine_googleapis_com_request_log_*`
WHERE
_TABLE_SUFFIX BETWEEN '20170901' AND '20170930')