Boost `1.88.0` introduced a feature [^1] that makes use of the Windows API, but it
uses API functions that are only available with `PSAPI_VERSION >= 2` and
Windows VISTA only supports `PSAPI_VERSION == 1`. Actually, that new feature
can also be disabled by setting the `BOOST_STACKTRACE_DISABLE_OFFSET_ADDR_BASE`
macro, but since it seems to be a useful feature and isn't even disabled by default,
we can just drop it that ancient Windows version instead of disabling it.
[^1]: https://github.com/boostorg/stacktrace/pull/200
With this mutex, we can have deadlock in the following case:
1/ Thread A processes a /v1/actions/acknowledge-problem request and locks the checkable
2/ Thread B processes a /v1/actions/add-comment and enters first the ConfigItem::ActivateItems() method and locks the static mutex there and starts the just created comment object, which triggers the OnCommentAdded() event.
3/ Thread A wants to activate the just created ack comment as well but since the mutex is already locked by TB, it blocks.
4/ Thread B's OnCommentAdded() even dispatch causes the IcingaDB::CommentAddedHandler() to be called and implicitly triggers full state update for the checkable. Now, the state serialization of that checkable (remember that's the same checkable currently locked by TA) also includes computing its severity, thus it calls either service->GetSeverity() or host->GetSeverity(). However, since computing the checkable severity (as of now) requires acquiring the object lock, and boom - they deadlock each other.
Pin child objects of hosts (HOST!...) to the same endpoint as the host.
This reduces cross-object action latency withing the same host.
If all endpoints know this algorithm, we can use it.
Old versions of OpenSSL stored a valid flag in the certificate (see inline code
comment for details) that if already set, causes parts of the verification to
be skipped and return that the certificate is valid, even if it's not actually
signed by the CA in the trust store.
This issue was assigned CVE-2025-48057.
`X509_STORE_CTX_get_error(csc)` was called after `X509_STORE_CTX_free(csc)`.
This is fixed by automatically freeing variables at the end of the function
using `std::unique_ptr`.
Previously, the `command#timeout` which by default is `1m`, was used to reschedule
the just sent remote check. However, this results into a bunch of extra checks being
sent to the remote host, even though the first one is still running. That's because
if one want to override the default timeout of the command for a specific host/service,
one has to set the `checkable#check_timeout` attribute to the desired value. So, this
commit makes sure that the `checkable#check_timeout` attribute (if set) is used to
reschedule the remote check.
Instead of IoEngine::YieldCurrentCoroutine(yc) until m_Queues.FutureResponseActions.empty(), async-wait a CV which is updated along with m_Queues.FutureResponseActions.
They're just useless, since a `CheckResult` handler is never going to be
called without a check result and a checkable can't exist without a
checkcommand.
Previously, recovery and ACK notifications were not delivered to users
who weren't notified about the problem state while having a configured
`Problem` type filter. However, since the type filter can also be
configured on the `Notification` object level, this resulted to an
incorrect behaviour. This PR changes the existing logic so that the
recovery and ACK notifications gets dropped only if the `Problem` filter
is configured on both the `User` and `Notification` object levels.
Ensure that the counter unit of measurement, "c", is parsed correctly
for performance data values again.
A prior refactoring in 720a88c29a489cec91815af49755413202802d7a changed
the parsing logic, resulting in an incorrect behavior for counter units.
By passing the raw input into the l_CsUoMs map first, the "c" UoM is
removed. Moving the explicit counter check before passing the raw unit
into the map resolves this issue.
Fixes#9540.
Previously, if you enable flapping for a Checkable but the corresponding
`Notification` object does not have `FlappingStart` or `FlappingEnd`
types set, the `no_more_notifications` flag wasn't reset to false again.
This commit ensures that this flag is always reset on `Recovery` even
the type filter does not match including when we miss the `Recovery` due
to Flapping state.
Notification objects can refer User and Group objects similar to how they can
refer Host and Service objects, so that dependency feels quite natural. Note
that for evaluating most configuration, this order doesn't really matter, the
configuration will successfully evaluate in either case, the difference can be
noticed mainly in more advanced configurations, for example when dynamically
assigning user based on their groups. When accessing user objects from the
Notification object definition (like in the following example), without this
change, only groups configured directly in groups attribute of User objects are
visible and those added via assign clauses in UserGroup objects are missing.
With this commit, these are also visible.
apply Notification "n" to Host {
for (var u in get_objects(User)) {
log(u.name + " -> " + Json.encode(u.groups))
}
# [...]
}
