ettore.dreucci/mk_livestatus
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
35b675e0392f4dcaaca4951490f1a04e97ea884f
mk_livestatus/src/TableStateHistory.cc
Phillip Smith 69e42aac8d initial commit
2021-08-24 17:55:37 +10:00

913 lines
36 KiB
C++
Raw Blame History

// +------------------------------------------------------------------+
// | ____ _ _ __ __ _ __ |
// | / ___| |__ ___ ___| | __ | \/ | |/ / |
// | | | | '_ \ / _ \/ __| |/ / | |\/| | ' / |
// | | |___| | | | __/ (__| < | | | | . \ |
// | \____|_| |_|\___|\___|_|\_\___|_| |_|_|\_\ |
// | |
// | Copyright Mathias Kettner 2014 mk@mathias-kettner.de |
// +------------------------------------------------------------------+
//
// This file is part of Check_MK.
// The official homepage is at http://mathias-kettner.de/check_mk.
//
// check_mk is free software; you can redistribute it and/or modify it
// under the terms of the GNU General Public License as published by
// the Free Software Foundation in version 2. check_mk is distributed
// in the hope that it will be useful, but WITHOUT ANY WARRANTY; with-
// out even the implied warranty of MERCHANTABILITY or FITNESS FOR A
// PARTICULAR PURPOSE. See the GNU General Public License for more de-
// tails. You should have received a copy of the GNU General Public
// License along with GNU Make; see the file COPYING. If not, write
// to the Free Software Foundation, Inc., 51 Franklin St, Fifth Floor,
// Boston, MA 02110-1301 USA.
#include "TableStateHistory.h"
#include <cstdint>
#include <ctime>
#include <memory>
#include <mutex>
#include <optional>
#include <ostream>
#include <set>
#include <stdexcept>
#include <utility>
#include <vector>
#include "Column.h"
#include "Filter.h"
#include "HostServiceState.h"
#include "LogEntry.h"
#include "Logger.h"
#include "OffsetDoubleColumn.h"
#include "OffsetIntColumn.h"
#include "OffsetSStringColumn.h"
#include "OffsetStringColumn.h"
#include "OffsetTimeColumn.h"
#include "OringFilter.h" // IWYU pragma: keep
#include "Query.h"
#include "Row.h"
#include "StringUtils.h"
#include "TableHosts.h"
#include "TableServices.h"
#ifdef CMC
// This seems to be an IWYU bug: If we remove the includes as suggested, we
// would do a member access on an incomplete type.
#include "Host.h" // IWYU pragma: keep
#include "Service.h" // IWYU pragma: keep
#include "Timeperiod.h"
#include "cmc.h"
#define STATE_OK 0
#define STATE_WARNING 1
#define STATE_CRITICAL 2
#define STATE_UNKNOWN 3
#else
#include "auth.h"
#include "nagios.h"
#endif
namespace {
constexpr unsigned classmask_statehist =
(1U << static_cast<int>(LogEntry::Class::alert)) | //
(1U << static_cast<int>(LogEntry::Class::program)) | //
(1U << static_cast<int>(LogEntry::Class::state)) | //
(1U << static_cast<int>(LogEntry::Class::text));
} // namespace
#ifndef CMC
namespace {
std::string getCustomVariable(customvariablesmember *cvm,
const std::string &name) {
for (; cvm != nullptr; cvm = cvm->next) {
if (cvm->variable_name == name) {
return cvm->variable_value == nullptr ? "" : cvm->variable_value;
}
}
return "";
}
} // namespace
#endif
TableStateHistory::TableStateHistory(MonitoringCore *mc, LogCache *log_cache)
: Table(mc), _log_cache(log_cache) {
addColumn(std::make_unique<OffsetTimeColumn>(
"time", "Time of the log event (seconds since 1/1/1970)", -1, -1, -1,
DANGEROUS_OFFSETOF(HostServiceState, _time)));
addColumn(std::make_unique<OffsetIntColumn>(
"lineno", "The number of the line in the log file", -1, -1, -1,
DANGEROUS_OFFSETOF(HostServiceState, _lineno)));
addColumn(std::make_unique<OffsetTimeColumn>(
"from", "Start time of state (seconds since 1/1/1970)", -1, -1, -1,
DANGEROUS_OFFSETOF(HostServiceState, _from)));
addColumn(std::make_unique<OffsetTimeColumn>(
"until", "End time of state (seconds since 1/1/1970)", -1, -1, -1,
DANGEROUS_OFFSETOF(HostServiceState, _until)));
addColumn(std::make_unique<OffsetIntColumn>(
"duration", "Duration of state (until - from)", -1, -1, -1,
DANGEROUS_OFFSETOF(HostServiceState, _duration)));
addColumn(std::make_unique<OffsetDoubleColumn>(
"duration_part", "Duration part in regard to the query timeframe", -1,
-1, -1, DANGEROUS_OFFSETOF(HostServiceState, _duration_part)));
addColumn(std::make_unique<OffsetIntColumn>(
"state",
"The state of the host or service in question - OK(0) / WARNING(1) / CRITICAL(2) / UNKNOWN(3) / UNMONITORED(-1)",
-1, -1, -1, DANGEROUS_OFFSETOF(HostServiceState, _state)));
addColumn(std::make_unique<OffsetIntColumn>(
"host_down", "Shows if the host of this service is down", -1, -1, -1,
DANGEROUS_OFFSETOF(HostServiceState, _host_down)));
addColumn(std::make_unique<OffsetIntColumn>(
"in_downtime", "Shows if the host or service is in downtime", -1, -1,
-1, DANGEROUS_OFFSETOF(HostServiceState, _in_downtime)));
addColumn(std::make_unique<OffsetIntColumn>(
"in_host_downtime", "Shows if the host of this service is in downtime",
-1, -1, -1, DANGEROUS_OFFSETOF(HostServiceState, _in_host_downtime)));
addColumn(std::make_unique<OffsetIntColumn>(
"is_flapping", "Shows if the host or service is flapping", -1, -1, -1,
DANGEROUS_OFFSETOF(HostServiceState, _is_flapping)));
addColumn(std::make_unique<OffsetIntColumn>(
"in_notification_period",
"Shows if the host or service is within its notification period", -1,
-1, -1, DANGEROUS_OFFSETOF(HostServiceState, _in_notification_period)));
addColumn(std::make_unique<OffsetStringColumn>(
"notification_period",
"The notification period of the host or service in question", -1, -1,
-1, DANGEROUS_OFFSETOF(HostServiceState, _notification_period)));
addColumn(std::make_unique<OffsetIntColumn>(
"in_service_period",
"Shows if the host or service is within its service period", -1, -1, -1,
DANGEROUS_OFFSETOF(HostServiceState, _in_service_period)));
addColumn(std::make_unique<OffsetSStringColumn>(
"service_period",
"The service period of the host or service in question", -1, -1, -1,
DANGEROUS_OFFSETOF(HostServiceState, _service_period)));
addColumn(std::make_unique<OffsetSStringColumn>(
"debug_info", "Debug information", -1, -1, -1,
DANGEROUS_OFFSETOF(HostServiceState, _debug_info)));
addColumn(std::make_unique<OffsetSStringColumn>(
"host_name", "Host name", -1, -1, -1,
DANGEROUS_OFFSETOF(HostServiceState, _host_name)));
addColumn(std::make_unique<OffsetSStringColumn>(
"service_description", "Description of the service", -1, -1, -1,
DANGEROUS_OFFSETOF(HostServiceState, _service_description)));
addColumn(std::make_unique<OffsetSStringColumn>(
"log_output", "Logfile output relevant for this state", -1, -1, -1,
DANGEROUS_OFFSETOF(HostServiceState, _log_output)));
addColumn(std::make_unique<OffsetIntColumn>(
"duration_ok", "OK duration of state ( until - from )", -1, -1, -1,
DANGEROUS_OFFSETOF(HostServiceState, _duration_state_OK)));
addColumn(std::make_unique<OffsetDoubleColumn>(
"duration_part_ok", "OK duration part in regard to the query timeframe",
-1, -1, -1, DANGEROUS_OFFSETOF(HostServiceState, _duration_part_OK)));
addColumn(std::make_unique<OffsetIntColumn>(
"duration_warning", "WARNING duration of state (until - from)", -1, -1,
-1, DANGEROUS_OFFSETOF(HostServiceState, _duration_state_WARNING)));
addColumn(std::make_unique<OffsetDoubleColumn>(
"duration_part_warning",
"WARNING duration part in regard to the query timeframe", -1, -1, -1,
DANGEROUS_OFFSETOF(HostServiceState, _duration_part_WARNING)));
addColumn(std::make_unique<OffsetIntColumn>(
"duration_critical", "CRITICAL duration of state (until - from)", -1,
-1, -1,
DANGEROUS_OFFSETOF(HostServiceState, _duration_state_CRITICAL)));
addColumn(std::make_unique<OffsetDoubleColumn>(
"duration_part_critical",
"CRITICAL duration part in regard to the query timeframe", -1, -1, -1,
DANGEROUS_OFFSETOF(HostServiceState, _duration_part_CRITICAL)));
addColumn(std::make_unique<OffsetIntColumn>(
"duration_unknown", "UNKNOWN duration of state (until - from)", -1, -1,
-1, DANGEROUS_OFFSETOF(HostServiceState, _duration_state_UNKNOWN)));
addColumn(std::make_unique<OffsetDoubleColumn>(
"duration_part_unknown",
"UNKNOWN duration part in regard to the query timeframe", -1, -1, -1,
DANGEROUS_OFFSETOF(HostServiceState, _duration_part_UNKNOWN)));
addColumn(std::make_unique<OffsetIntColumn>(
"duration_unmonitored", "UNMONITORED duration of state (until - from)",
-1, -1, -1,
DANGEROUS_OFFSETOF(HostServiceState, _duration_state_UNMONITORED)));
addColumn(std::make_unique<OffsetDoubleColumn>(
"duration_part_unmonitored",
"UNMONITORED duration part in regard to the query timeframe", -1, -1,
-1, DANGEROUS_OFFSETOF(HostServiceState, _duration_part_UNMONITORED)));
// join host and service tables
TableHosts::addColumns(this, "current_host_",
DANGEROUS_OFFSETOF(HostServiceState, _host), -1);
TableServices::addColumns(this, "current_service_",
DANGEROUS_OFFSETOF(HostServiceState, _service),
false /* no hosts table */);
}
std::string TableStateHistory::name() const { return "statehist"; }
std::string TableStateHistory::namePrefix() const { return "statehist_"; }
void TableStateHistory::getPreviousLogentry() {
while (_it_entries == _entries->begin()) {
// open previous logfile
if (_it_logs == _log_cache->begin()) {
return;
}
--_it_logs;
_entries = _it_logs->second->getEntriesFor(classmask_statehist);
_it_entries = _entries->end();
}
--_it_entries;
}
LogEntry *TableStateHistory::getNextLogentry() {
if (_it_entries != _entries->end()) {
++_it_entries;
}
while (_it_entries == _entries->end()) {
auto it_logs_cpy = _it_logs;
if (++it_logs_cpy == _log_cache->end()) {
return nullptr;
}
++_it_logs;
_entries = _it_logs->second->getEntriesFor(classmask_statehist);
_it_entries = _entries->begin();
}
return _it_entries->second.get();
}
namespace {
class TimeperiodTransition {
public:
explicit TimeperiodTransition(const std::string &str) {
auto fields = mk::split(str, ';');
if (fields.size() != 3) {
throw std::invalid_argument("expected 3 arguments");
}
_name = fields[0];
_from = std::stoi(fields[1]);
_to = std::stoi(fields[2]);
}
[[nodiscard]] std::string name() const { return _name; }
[[nodiscard]] int from() const { return _from; }
[[nodiscard]] int to() const { return _to; }
private:
std::string _name;
int _from;
int _to;
};
} // namespace
// Create a partial filter, that contains only such filters that check
// attributes of current hosts and services
// static
std::unique_ptr<Filter> TableStateHistory::createPartialFilter(
const Query &query) {
return query.partialFilter(
"current host/service columns", [](const Column &column) {
return mk::starts_with(column.name(), "current_") ||
mk::starts_with(column.name(), "host_") ||
mk::starts_with(column.name(), "service_");
});
}
void TableStateHistory::answerQuery(Query *query) {
auto object_filter = createPartialFilter(*query);
std::lock_guard<std::mutex> lg(_log_cache->_lock);
_log_cache->update();
if (_log_cache->begin() == _log_cache->end()) {
return;
}
// This flag might be set to true by the return value of processDataset(...)
_abort_query = false;
// Keep track of the historic state of services/hosts here
std::map<HostServiceKey, HostServiceState *> state_info;
// Store hosts/services that we have filtered out here
std::set<HostServiceKey> object_blacklist;
// Optimize time interval for the query. In log querys there should always
// be a time range in form of one or two filter expressions over time. We
// use that to limit the number of logfiles we need to scan and to find the
// optimal entry point into the logfile
if (auto glb = query->greatestLowerBoundFor("time")) {
_since = *glb;
} else {
query->invalidRequest(
"Start of timeframe required. e.g. Filter: time > 1234567890");
return;
}
_until = query->leastUpperBoundFor("time").value_or(time(nullptr)) + 1;
_query_timeframe = _until - _since - 1;
if (_query_timeframe == 0) {
query->invalidRequest("Query timeframe is 0 seconds");
return;
}
// Switch to last logfile (we have at least one)
_it_logs = _log_cache->end();
--_it_logs;
auto newest_log = _it_logs;
// Now find the log where 'since' starts.
while (_it_logs != _log_cache->begin() && _it_logs->first >= _since) {
--_it_logs; // go back in history
}
// Check if 'until' is within these logfiles
if (_it_logs->first > _until) {
// All logfiles are too new, invalid timeframe
// -> No data available. Return empty result.
return;
}
// Determine initial logentry
_entries = _it_logs->second->getEntriesFor(classmask_statehist);
if (!_entries->empty() && _it_logs != newest_log) {
_it_entries = _entries->end();
// Check last entry. If it's younger than _since -> use this logfile too
if (--_it_entries != _entries->begin()) {
if (_it_entries->second->_time >= _since) {
_it_entries = _entries->begin();
}
}
} else {
_it_entries = _entries->begin();
}
// From now on use getPreviousLogentry() / getNextLogentry()
bool only_update = true;
bool in_nagios_initial_states = false;
while (LogEntry *entry = getNextLogentry()) {
if (_abort_query) {
break;
}
if (entry->_time >= _until) {
getPreviousLogentry();
break;
}
if (only_update && entry->_time >= _since) {
// Reached start of query timeframe. From now on let's produce real
// output. Update _from time of every state entry
for (auto &it_hst : state_info) {
it_hst.second->_from = _since;
it_hst.second->_until = _since;
}
only_update = false;
}
if (in_nagios_initial_states &&
!(entry->_type == LogEntryType::state_service_initial ||
entry->_type == LogEntryType::state_host_initial)) {
// Set still unknown hosts / services to unmonitored
for (auto &it_hst : state_info) {
HostServiceState *hst = it_hst.second;
if (hst->_may_no_longer_exist) {
hst->_has_vanished = true;
}
}
in_nagios_initial_states = false;
}
HostServiceKey key = nullptr;
bool is_service = false;
switch (entry->_type) {
case LogEntryType::none:
case LogEntryType::core_starting:
case LogEntryType::core_stopping:
case LogEntryType::log_version:
case LogEntryType::acknowledge_alert_host:
case LogEntryType::acknowledge_alert_service:
break;
case LogEntryType::alert_service:
case LogEntryType::state_service:
case LogEntryType::state_service_initial:
case LogEntryType::downtime_alert_service:
case LogEntryType::flapping_service:
key = entry->_service;
is_service = true;
// fall-through
case LogEntryType::alert_host:
case LogEntryType::state_host:
case LogEntryType::state_host_initial:
case LogEntryType::downtime_alert_host:
case LogEntryType::flapping_host: {
if (!is_service) {
key = entry->_host;
}
if (key == nullptr) {
continue;
}
if (object_blacklist.find(key) != object_blacklist.end()) {
// Host/Service is not needed for this query and has already
// been filtered out.
continue;
}
// Find state object for this host/service
HostServiceState *state;
auto it_hst = state_info.find(key);
if (it_hst == state_info.end()) {
// Create state object that we also need for filtering right
// now
state = new HostServiceState();
state->_is_host = entry->_svc_desc.empty();
state->_host = entry->_host;
state->_service = entry->_service;
#ifdef CMC
state->_host_name = entry->_host->name();
state->_service_description = entry->_service == nullptr
? ""
: entry->_service->name();
#else
state->_host_name = entry->_host->name;
state->_service_description =
entry->_service == nullptr
? ""
: entry->_service->description;
#endif
// No state found. Now check if this host/services is
// filtered out. Note: we currently do not filter out hosts
// since they might be needed for service states
if (!entry->_svc_desc.empty()) {
if (!object_filter->accepts(Row(state),
query->authUser(),
query->timezoneOffset())) {
object_blacklist.insert(key);
delete state;
continue;
}
}
// Host/Service relations
if (state->_is_host) {
for (auto &it_inh : state_info) {
if (it_inh.second->_host == state->_host) {
state->_services.push_back(it_inh.second);
}
}
} else {
auto it_inh = state_info.find(state->_host);
if (it_inh != state_info.end()) {
it_inh->second->_services.push_back(state);
}
}
// Store this state object for tracking state transitions
state_info.emplace(key, state);
state->_from = _since;
// Get notification period of host/service
// If this host/service is no longer availabe in nagios ->
// set to ""
if (state->_service != nullptr) {
#ifdef CMC
state->_notification_period =
state->_service->notificationPeriod()->name();
#else
auto np = state->_service->notification_period;
state->_notification_period = np == nullptr ? "" : np;
#endif
} else if (state->_host != nullptr) {
#ifdef CMC
state->_notification_period =
state->_host->notificationPeriod()->name();
#else
auto np = state->_host->notification_period;
state->_notification_period = np == nullptr ? "" : np;
#endif
} else {
state->_notification_period = "";
}
// Same for service period. For Nagios this is a bit
// different, since this is no native field but just a
// custom variable
if (state->_service != nullptr) {
#ifdef CMC
state->_service_period =
state->_service->servicePeriod()->name();
#else
state->_service_period =
getCustomVariable(state->_service->custom_variables,
"SERVICE_PERIOD");
#endif
} else if (state->_host != nullptr) {
#ifdef CMC
state->_service_period =
state->_host->servicePeriod()->name();
#else
state->_service_period = getCustomVariable(
state->_host->custom_variables, "SERVICE_PERIOD");
#endif
} else {
state->_service_period = "";
}
// Determine initial in_notification_period status
auto tmp_period =
_notification_periods.find(state->_notification_period);
if (tmp_period != _notification_periods.end()) {
state->_in_notification_period = tmp_period->second;
} else {
state->_in_notification_period = 1;
}
// Same for service period
tmp_period =
_notification_periods.find(state->_service_period);
if (tmp_period != _notification_periods.end()) {
state->_in_service_period = tmp_period->second;
} else {
state->_in_service_period = 1;
}
// If this key is a service try to find its host and apply
// its _in_host_downtime and _host_down parameters
if (!state->_is_host) {
auto my_host = state_info.find(state->_host);
if (my_host != state_info.end()) {
state->_in_host_downtime =
my_host->second->_in_host_downtime;
state->_host_down = my_host->second->_host_down;
}
}
// Log UNMONITORED state if this host or service just
// appeared within the query timeframe
// It gets a grace period of ten minutes (nagios startup)
if (!only_update && entry->_time - _since > 60 * 10) {
state->_debug_info = "UNMONITORED ";
state->_state = -1;
}
} else {
state = it_hst->second;
}
int state_changed =
updateHostServiceState(query, entry, state, only_update);
// Host downtime or state changes also affect its services
if (entry->_type == LogEntryType::alert_host ||
entry->_type == LogEntryType::state_host ||
entry->_type == LogEntryType::downtime_alert_host) {
if (state_changed != 0) {
for (auto &_service : state->_services) {
updateHostServiceState(query, entry, _service,
only_update);
}
}
}
break;
}
case LogEntryType::timeperiod_transition: {
try {
TimeperiodTransition tpt(entry->_options);
_notification_periods[tpt.name()] = tpt.to();
for (auto &it_hst : state_info) {
updateHostServiceState(query, entry, it_hst.second,
only_update);
}
} catch (const std::logic_error &e) {
Warning(logger())
<< "Error: Invalid syntax of TIMEPERIOD TRANSITION: "
<< entry->_complete;
}
break;
}
case LogEntryType::log_initial_states: {
// This feature is only available if log_initial_states is set
// to 1. If log_initial_states is set, each nagios startup logs
// the initial states of all known hosts and services. Therefore
// we can detect if a host is no longer available after a nagios
// startup. If it still exists an INITIAL HOST/SERVICE state
// entry will follow up shortly.
for (auto &it_hst : state_info) {
if (!it_hst.second->_has_vanished) {
it_hst.second->_last_known_time = entry->_time;
it_hst.second->_may_no_longer_exist = true;
}
}
in_nagios_initial_states = true;
break;
}
}
}
// Create final reports
auto it_hst = state_info.begin();
if (!_abort_query) {
while (it_hst != state_info.end()) {
HostServiceState *hst = it_hst->second;
// No trace since the last two nagios startup -> host/service has
// vanished
if (hst->_may_no_longer_exist) {
// Log last known state up to nagios restart
hst->_time = hst->_last_known_time;
hst->_until = hst->_last_known_time;
process(query, hst);
// Set absent state
hst->_state = -1;
hst->_debug_info = "UNMONITORED";
hst->_log_output = "";
}
hst->_time = _until - 1;
hst->_until = hst->_time;
process(query, hst);
++it_hst;
}
}
// Cleanup !
it_hst = state_info.begin();
while (it_hst != state_info.end()) {
delete it_hst->second;
++it_hst;
}
state_info.clear();
object_blacklist.clear();
}
int TableStateHistory::updateHostServiceState(Query *query,
const LogEntry *entry,
HostServiceState *hs_state,
const bool only_update) {
int state_changed = 1;
// Revive host / service if it was unmonitored
if (entry->_type != LogEntryType::timeperiod_transition &&
hs_state->_has_vanished) {
hs_state->_time = hs_state->_last_known_time;
hs_state->_until = hs_state->_last_known_time;
if (!only_update) {
process(query, hs_state);
}
hs_state->_may_no_longer_exist = false;
hs_state->_has_vanished = false;
// Set absent state
hs_state->_state = -1;
hs_state->_debug_info = "UNMONITORED";
hs_state->_in_downtime = 0;
hs_state->_in_notification_period = 0;
hs_state->_in_service_period = 0;
hs_state->_is_flapping = 0;
hs_state->_log_output = "";
// Apply latest notification period information and set the host_state
// to unmonitored
auto it_status =
_notification_periods.find(hs_state->_notification_period);
if (it_status != _notification_periods.end()) {
hs_state->_in_notification_period = it_status->second;
} else {
// No notification period information available -> within
// notification period
hs_state->_in_notification_period = 1;
}
// Same for service period
it_status = _notification_periods.find(hs_state->_service_period);
if (it_status != _notification_periods.end()) {
hs_state->_in_service_period = it_status->second;
} else {
// No service period information available -> within service period
hs_state->_in_service_period = 1;
}
}
// Update basic information
hs_state->_time = entry->_time;
hs_state->_lineno = entry->_lineno;
hs_state->_until = entry->_time;
// A timeperiod entry never brings an absent host or service into
// existence..
if (entry->_type != LogEntryType::timeperiod_transition) {
hs_state->_may_no_longer_exist = false;
}
switch (entry->_type) {
case LogEntryType::none:
case LogEntryType::core_starting:
case LogEntryType::core_stopping:
case LogEntryType::log_version:
case LogEntryType::log_initial_states:
case LogEntryType::acknowledge_alert_host:
case LogEntryType::acknowledge_alert_service:
break;
case LogEntryType::state_host:
case LogEntryType::state_host_initial:
case LogEntryType::alert_host: {
if (hs_state->_is_host) {
if (hs_state->_state != entry->_state) {
if (!only_update) {
process(query, hs_state);
}
hs_state->_state = entry->_state;
hs_state->_host_down = static_cast<int>(entry->_state > 0);
hs_state->_debug_info = "HOST STATE";
} else {
state_changed = 0;
}
} else if (hs_state->_host_down !=
static_cast<int>(entry->_state > 0)) {
if (!only_update) {
process(query, hs_state);
}
hs_state->_host_down = static_cast<int>(entry->_state > 0);
hs_state->_debug_info = "SVC HOST STATE";
}
break;
}
case LogEntryType::state_service:
case LogEntryType::state_service_initial:
case LogEntryType::alert_service: {
if (hs_state->_state != entry->_state) {
if (!only_update) {
process(query, hs_state);
}
hs_state->_debug_info = "SVC ALERT";
hs_state->_state = entry->_state;
}
break;
}
case LogEntryType::downtime_alert_host: {
int downtime_active =
mk::starts_with(entry->_state_type, "STARTED") ? 1 : 0;
if (hs_state->_in_host_downtime != downtime_active) {
if (!only_update) {
process(query, hs_state);
}
hs_state->_debug_info =
hs_state->_is_host ? "HOST DOWNTIME" : "SVC HOST DOWNTIME";
hs_state->_in_host_downtime = downtime_active;
if (hs_state->_is_host) {
hs_state->_in_downtime = downtime_active;
}
} else {
state_changed = 0;
}
break;
}
case LogEntryType::downtime_alert_service: {
int downtime_active =
mk::starts_with(entry->_state_type, "STARTED") ? 1 : 0;
if (hs_state->_in_downtime != downtime_active) {
if (!only_update) {
process(query, hs_state);
}
hs_state->_debug_info = "DOWNTIME SERVICE";
hs_state->_in_downtime = downtime_active;
}
break;
}
case LogEntryType::flapping_host:
case LogEntryType::flapping_service: {
int flapping_active =
mk::starts_with(entry->_state_type, "STARTED") ? 1 : 0;
if (hs_state->_is_flapping != flapping_active) {
if (!only_update) {
process(query, hs_state);
}
hs_state->_debug_info = "FLAPPING ";
hs_state->_is_flapping = flapping_active;
} else {
state_changed = 0;
}
break;
}
case LogEntryType::timeperiod_transition: {
try {
TimeperiodTransition tpt(entry->_options);
// if no _host pointer is available the initial status of
// _in_notification_period (1) never changes
if (hs_state->_host != nullptr &&
tpt.name() == hs_state->_notification_period) {
if (tpt.to() != hs_state->_in_notification_period) {
if (!only_update) {
process(query, hs_state);
}
hs_state->_debug_info = "TIMEPERIOD ";
hs_state->_in_notification_period = tpt.to();
}
}
// same for service period
if (hs_state->_host != nullptr &&
tpt.name() == hs_state->_service_period) {
if (tpt.to() != hs_state->_in_service_period) {
if (!only_update) {
process(query, hs_state);
}
hs_state->_debug_info = "TIMEPERIOD ";
hs_state->_in_service_period = tpt.to();
}
}
} catch (const std::logic_error &e) {
Warning(logger())
<< "Error: Invalid syntax of TIMEPERIOD TRANSITION: "
<< entry->_complete;
}
break;
}
}
if (entry->_type != LogEntryType::timeperiod_transition) {
if ((entry->_type == LogEntryType::state_host_initial ||
entry->_type == LogEntryType::state_service_initial) &&
entry->_check_output == "(null)") {
hs_state->_log_output = "";
} else {
hs_state->_log_output = entry->_check_output;
}
}
return state_changed;
}
void TableStateHistory::process(Query *query, HostServiceState *hs_state) {
hs_state->_duration = hs_state->_until - hs_state->_from;
hs_state->_duration_part = static_cast<double>(hs_state->_duration) /
static_cast<double>(_query_timeframe);
hs_state->_duration_state_UNMONITORED = 0;
hs_state->_duration_part_UNMONITORED = 0;
hs_state->_duration_state_OK = 0;
hs_state->_duration_part_OK = 0;
hs_state->_duration_state_WARNING = 0;
hs_state->_duration_part_WARNING = 0;
hs_state->_duration_state_CRITICAL = 0;
hs_state->_duration_part_CRITICAL = 0;
hs_state->_duration_state_UNKNOWN = 0;
hs_state->_duration_part_UNKNOWN = 0;
switch (hs_state->_state) {
case -1:
hs_state->_duration_state_UNMONITORED = hs_state->_duration;
hs_state->_duration_part_UNMONITORED = hs_state->_duration_part;
break;
case STATE_OK:
hs_state->_duration_state_OK = hs_state->_duration;
hs_state->_duration_part_OK = hs_state->_duration_part;
break;
case STATE_WARNING:
hs_state->_duration_state_WARNING = hs_state->_duration;
hs_state->_duration_part_WARNING = hs_state->_duration_part;
break;
case STATE_CRITICAL:
hs_state->_duration_state_CRITICAL = hs_state->_duration;
hs_state->_duration_part_CRITICAL = hs_state->_duration_part;
break;
case STATE_UNKNOWN:
hs_state->_duration_state_UNKNOWN = hs_state->_duration;
hs_state->_duration_part_UNKNOWN = hs_state->_duration_part;
break;
default:
break;
}
// if (hs_state->_duration > 0)
_abort_query = !query->processDataset(Row(hs_state));
hs_state->_from = hs_state->_until;
}
bool TableStateHistory::isAuthorized(Row row, const contact *ctc) const {
auto entry = rowData<HostServiceState>(row);
service *svc = entry->_service;
host *hst = entry->_host;
return (hst != nullptr || svc != nullptr) &&
is_authorized_for(core(), ctc, hst, svc);
}
std::shared_ptr<Column> TableStateHistory::column(std::string colname) const {
try {
// First try to find column in the usual way
return Table::column(colname);
} catch (const std::runtime_error &e) {
// Now try with prefix "current_", since our joined tables have this
// prefix in order to make clear that we access current and not historic
// data and in order to prevent mixing up historic and current fields
// with the same name.
return Table::column("current_" + colname);
}
}
Reference in New Issue View Git Blame Copy Permalink