def get_check_table(discovery_request: StartDiscoveryRequest) -> DiscoveryResult:
"""Gathers the check table using a background job
Cares about handling local / remote sites using an automation call. In both cases
the ServiceDiscoveryBackgroundJob is executed to care about collecting the check
table asynchronously. In case of a remote site the chain is:
Starting from central site:
_get_check_table()
|
v
automation service-discovery-job-discover
|
v
to remote site
|
v
AutomationServiceDiscoveryJob().execute()
|
v
_get_check_table()
"""
if discovery_request.options.action == DiscoveryAction.TABULA_RASA:
watolib.add_service_change(
discovery_request.host, "refresh-autochecks",
_("Refreshed check configuration of host '%s'") % discovery_request.host.name())
if config.site_is_local(discovery_request.host.site_id()):
return execute_discovery_job(discovery_request)
discovery_result = _get_check_table_from_remote(discovery_request)
discovery_result = _add_missing_discovery_result_fields(discovery_result)
return discovery_result
def _discover_services(self, request):
mode = request.get("mode", "new")
hostname = request.get("hostname")
check_hostname(hostname, should_exist=True)
host = watolib.Host.host(hostname)
host_attributes = host.effective_attributes()
if host.is_cluster():
# This is currently the only way to get some actual discovery statitics.
# Start a dry-run -> Get statistics
# Do an actual discovery on the nodes -> data is written
try_result = watolib.check_mk_automation(host_attributes.get("site"), "try-inventory",
["@scan"] + [hostname])
new = 0
old = 0
for entry in try_result["check_table"]:
if entry[0] == "new":
new += 1
elif entry[0] == "old":
old += 1
result = DiscoveryResult(self_new=new, self_kept=old, self_total=new + old)
watolib.check_mk_automation(host_attributes.get("site"), "inventory",
["@scan", mode] + host.cluster_nodes())
else:
response = execute_automation_discovery(site_id=host_attributes.get("site"),
args=["@scan", mode, hostname])
result = response.results[hostname]
if result.error_text:
if not host.discovery_failed():
host.set_discovery_failed()
raise MKUserError(None, _("Failed to discover %s: %s") % (hostname, result.error_text))
if host.discovery_failed():
host.clear_discovery_failed()
if mode == "refresh":
message = _("Refreshed check configuration of host [%s] with %d services") % (
hostname, result.self_total)
watolib.add_service_change(host, "refresh-autochecks", message)
else:
message = _("Saved check configuration of host [%s] with %d services") % (
hostname, result.self_total)
watolib.add_service_change(host, "set-autochecks", message)
msg = _("Service discovery successful. Added %d, removed %d, kept %d, total %d services "
"and %d new, %d total host labels") % (
result.self_new,
result.self_removed,
result.self_kept,
result.self_total,
result.self_new_host_labels,
result.self_total_host_labels,
)
return msg
def _discover_services(self, request):
mode = request.get("mode", "new")
hostname = request.get("hostname")
check_hostname(hostname, should_exist=True)
host = watolib.Host.host(hostname)
host_attributes = host.effective_attributes()
if host.is_cluster():
# This is currently the only way to get some actual discovery statitics.
# Start a dry-run -> Get statistics
# Do an actual discovery on the nodes -> data is written
result = watolib.check_mk_automation(host_attributes.get("site"), "try-inventory",
["@scan"] + [hostname])
counts = {"new": 0, "old": 0}
for entry in result["check_table"]:
if entry[0] in counts:
counts[entry[0]] += 1
counts = {
hostname: (
counts["new"],
0, # this info is not available for clusters
counts["old"],
counts["new"] + counts["old"])
}
# A cluster cannot fail, just the nodes. This information is currently discarded
failed_hosts = None
watolib.check_mk_automation(host_attributes.get("site"), "inventory",
["@scan", mode] + host.cluster_nodes())
else:
counts, failed_hosts = watolib.check_mk_automation(host_attributes.get("site"),
"inventory",
["@scan", mode] + [hostname])
if failed_hosts:
if not host.discovery_failed():
host.set_discovery_failed()
raise MKUserError(
None,
_("Failed to inventorize %s: %s") % (hostname, failed_hosts[hostname]))
if host.discovery_failed():
host.clear_discovery_failed()
if mode == "refresh":
message = _("Refreshed check configuration of host [%s] with %d services") % (
hostname, counts[hostname][3])
watolib.add_service_change(host, "refresh-autochecks", message)
else:
message = _("Saved check configuration of host [%s] with %d services") % (
hostname, counts[hostname][3])
watolib.add_service_change(host, "set-autochecks", message)
msg = _("Service discovery successful. Added %d, removed %d, kept %d, total %d services "
"and %d new, %d total host labels") % tuple(counts[hostname])
return msg
def _save_services(self, old_autochecks: SetAutochecksTable,
checks: SetAutochecksTable, need_sync: bool) -> None:
message = _(
"Saved check configuration of host '%s' with %d services") % (
self._host.name(),
len(checks),
)
watolib.add_service_change(
host=self._host,
action_name="set-autochecks",
text=message,
need_sync=need_sync,
diff_text=watolib.make_diff_text(
_make_host_audit_log_object(old_autochecks),
_make_host_audit_log_object(checks)),
)
site_id = self._host.site_id()
site_status = states().get(site_id, SiteStatus({}))
if is_pre_17_remote_site(site_status):
# is this branch still needed?
set_autochecks(
site_id,
self._host.name(),
{x: y[1:3]
for x, y in checks.items()}, # type: ignore[misc]
)
else:
set_autochecks(
site_id,
self._host.name(),
checks,
)
def _save_services(self, checks, need_sync):
message = _("Saved check configuration of host '%s' with %d services") % \
(self._host.name(), len(checks))
watolib.add_service_change(self._host,
"set-autochecks",
message,
need_sync=need_sync)
check_mk_automation(self._host.site_id(), "set-autochecks",
[self._host.name()], checks)
def _save_services(self, old_autochecks: SetAutochecksTable,
checks: SetAutochecksTable, need_sync: bool) -> None:
message = _("Saved check configuration of host '%s' with %d services"
) % (self._host.name(), len(checks))
watolib.add_service_change(
host=self._host,
action_name="set-autochecks",
text=message,
need_sync=need_sync,
diff_text=watolib.make_diff_text(
_make_host_audit_log_object(old_autochecks),
_make_host_audit_log_object(checks)),
)
check_mk_automation(self._host.site_id(), "set-autochecks",
[self._host.name()], checks)
def get_check_table(
discovery_request: StartDiscoveryRequest) -> DiscoveryResult:
"""Gathers the check table using a background job
Cares about handling local / remote sites using an automation call. In both cases
the ServiceDiscoveryBackgroundJob is executed to care about collecting the check
table asynchronously. In case of a remote site the chain is:
Starting from central site:
_get_check_table()
|
v
automation service-discovery-job-discover
|
v
to remote site
|
v
AutomationServiceDiscoveryJob().execute()
|
v
_get_check_table()
"""
if discovery_request.options.action == DiscoveryAction.TABULA_RASA:
watolib.add_service_change(
discovery_request.host,
"refresh-autochecks",
_("Refreshed check configuration of host '%s'") %
discovery_request.host.name(),
)
if site_is_local(discovery_request.host.site_id()):
return execute_discovery_job(discovery_request)
sync_changes_before_remote_automation(discovery_request.host.site_id())
return DiscoveryResult.deserialize(
watolib.do_remote_automation(
get_site_config(discovery_request.host.site_id()),
"service-discovery-job",
[
("host_name", discovery_request.host.name()),
("options", json.dumps(discovery_request.options._asdict())),
],
))
