6874 commits

Author	SHA1	Message	Date
Julian Brost	207764584a	Redis exceptions: remove inline specifiers ... Remove them as they are redundant, as requested in the PR review.	2026-03-27 17:02:05 +01:00
Julian Brost	221382486e	Redis exceptions: use BOOST_THROW_EXCEPTION ... Use it for consistency, as requested in the PR review.	2026-03-27 17:01:51 +01:00
Julian Brost	862f012381	Redis exceptions: add proper what() messages ... RedisDisconnected::what() and RedisProtocolError::what() always returned an empty string. Similarly, BadRedisType::what() and BadRedisInt::what() only return the value that couldn't be parsed without any information about the exception type. If only what() is used when printing the exception, as it's typical, this results in unhelpful log messages like the following when simply stopping the Redis server: [2026-02-23 14:33:33 +0100] critical/IcingaDB: Error during receiving the response to a query which has been fired and forgotten: Connection reset by peer [system:104 at /usr/include/boost/asio/detail/reactive_socket_recv_op.hpp:134 in function 'do_complete'] [2026-02-23 14:33:33 +0100] critical/IcingaDB: Error during receiving the response to a query which has been fired and forgotten: [2026-02-23 14:33:33 +0100] critical/IcingaDB: Error during receiving the response to a query which has been fired and forgotten: [2026-02-23 14:33:33 +0100] critical/IcingaDB: Cannot connect to redis-1:6379: Connection refused [system:111 at /usr/include/boost/asio/detail/reactive_socket_connect_op.hpp:98 in function 'do_complete'] This commit changes these messages so that something like "Redis disconnected", "Redis protocol error: bad int: foo", or "Redis protocol error: bad type: ?" is returned. In doing so, it also removes a member of type std::vector<char> in BadRedisInt as this is unsafe to use in exceptions (it violates the requirement that copy constructor and assignment must be nothrow, see https://en.cppreference.com/w/cpp/error/exception.html#Standard_exception_requirements). With this commit, the log messages are now a bit more helpful: [2026-02-23 15:08:23 +0100] critical/IcingaDB: Error during receiving the response to a query which has been fired and forgotten: Connection reset by peer [system:104 at /usr/include/boost/asio/detail/reactive_socket_recv_op.hpp:134 in function 'do_complete'] [2026-02-23 15:08:23 +0100] critical/IcingaDB: Error during receiving the response to a query which has been fired and forgotten: Redis disconnected [2026-02-23 15:08:23 +0100] critical/IcingaDB: Error during receiving the response to a query which has been fired and forgotten: Redis disconnected [2026-02-23 15:08:23 +0100] critical/IcingaDB: Cannot connect to redis-1:6379: Connection refused [system:111 at /usr/include/boost/asio/detail/reactive_socket_connect_op.hpp:98 in function 'do_complete']	2026-03-27 17:01:49 +01:00
Johannes Schmidt	2108300cf8	Add warnings to deprecated features indicating removal in v2.18	2026-03-27 14:20:55 +01:00
Yonas Habteab	895dea2dc4	Merge pull request #10770 from Icinga/problem-chars ... Warn on problematic object names	2026-03-27 10:21:58 +01:00
Yonas Habteab	4039b39061	Warn if object name contains ctrl characters	2026-03-26 15:01:15 +01:00
Yonas Habteab	da521203b1	Endpoint: warn if endpoint name exceeds 64 characters	2026-03-26 13:58:57 +01:00
Yonas Habteab	5a4cfae182	Warn of object name exceeds 255 characters	2026-03-26 13:58:57 +01:00
Yonas Habteab	2e6d23701a	Warn about leading and trailing WSP in object names	2026-03-26 13:58:57 +01:00
Julian Brost	ed403294a3	OutgoingHttpMessage: don't use shared_ptr for m_CpuBoundWork ... This change just gives clear ownership over the CpuBoundWork to the OutgoingHttpMessage, instead of the previous shared_ptr and weak_ptr combination with an unclear purpose.	2026-03-26 11:47:45 +01:00
Johannes Schmidt	7d4af75189	Make ShowCodeLocation() show the full line with a LastColumn of 0	2026-03-24 13:53:31 +01:00
Yonas Habteab	26d23bd569	IcingaDB: enqueue all dependency related updates to the worker queue	2026-03-23 11:26:48 +01:00
Yonas Habteab	e6dc594313	IcingaDB: don't send nextupdate:{host,service} as part of the initial sync ... Don't send nextupdate as part of the intial sync. Instead, enque them to the background worker to be sent after the initial dump is done.	2026-03-23 11:26:48 +01:00
Yonas Habteab	eab507f892	IcingaDB: track object changesets efficiently	2026-03-23 11:26:48 +01:00
Yonas Habteab	105c066649	IcingaDB: use a separate Redis conn for history & heartbeat streams	2026-03-23 11:26:47 +01:00
Alexander A. Klimov	85b16d2186	Remove unused IoEngine#m_AlreadyExpiredTimer	2026-03-19 14:53:29 +01:00
Alexander A. Klimov	7b45d8d625	Remove unused IoEngine::YieldCurrentCoroutine()	2026-03-19 14:53:29 +01:00
Alexander A. Klimov	cefa942c31	Make IoEngine#m_CpuBoundSemaphore unsigned	2026-03-19 14:53:29 +01:00
Alexander A. Klimov	06cfdbb650	CpuBoundWork#CpuBoundWork(): don't spin on atomic int to acquire slot ... This is inefficient and involves possible bad surprises regarding waiting durations on busy nodes. Instead, use `AsioConditionVariable#Wait()` if there are no free slots. It's notified by others' `CpuBoundWork#~CpuBoundWork()` once finished.	2026-03-19 14:53:29 +01:00
Alexander A. Klimov	c40c15ee11	Introduce AsioConditionVariable	2026-03-19 14:53:29 +01:00
Alexander A. Klimov	e03db5f71d	[Refactor] CpuBoundWork#CpuBoundWork(): require an io_context::strand	2026-03-19 14:53:29 +01:00
Alexander A. Klimov	149aecf3b9	CpuBoundWork#~CpuBoundWork(): just call #Done() to deduplicate code	2026-03-19 14:53:29 +01:00
Alexander A. Klimov	844b341225	Remove unused IoBoundWorkSlot	2026-03-19 14:53:29 +01:00
Alexander A. Klimov	fdc08c2e00	OutgoingHttpMessage#Flush(): release CpuBoundWork slot ... so that `/v1/events` doesn't have to use `IoBoundWorkSlot`. `IoBoundWorkSlot#~IoBoundWorkSlot()` will wait for a free semaphore slot which will be almost immediately released by `CpuBoundWork#~CpuBoundWork()`. Just releasing the already aquired slot manually is more efficient.	2026-03-19 14:51:19 +01:00
Julian Brost	6592eae21d	Merge pull request #10668 from Icinga/perfdata-writers-connection-handling ... Add PerfdatawriterConnection to handle network requests for Perfdata Writers	2026-03-17 15:10:57 +01:00
Johannes Schmidt	da2fc9df01	Use PerfdataWriterConnection in perfdata writers	2026-03-17 12:11:26 +01:00
Yonas Habteab	fa514e6336	IcingaDB: use a statically initialized list of syncable types for Redis keys	2026-03-16 08:39:12 +01:00
Yonas Habteab	cb872ee5c8	IcingaDB: consistently use const char* literals everywhere ... Co-Authored-By: Julian Brost <julian.brost@icinga.com>	2026-03-13 16:14:12 +01:00
Yonas Habteab	f593330735	IcingaDB: change const T to const T& to avoid stupid String copying	2026-03-13 16:14:12 +01:00
Johannes Schmidt	8f85e8b72a	Merge pull request #10756 from Icinga/improve-parallel-for ... Improve `WorkQueue::ParallelFor()`	2026-03-13 16:12:32 +01:00
Johannes Schmidt	907f41bae4	Use perfect-forwarding references to prevent passing temporaries ... Without this change, it is easy to pass a temporary, for example a vector returned by a function to ParallelFor(). With this commit a type-constraint is added to disable the use of this function for rvalue sequences. An alternative would have been to capture rvalue-references in the function, but that would have made the function unnecessarily complex, involving tuple-capture and `shared_ptr`s, while it is usually easy to control lifetime at the call-site.	2026-03-13 13:14:52 +01:00
Johannes Schmidt	eb97638743	Make retrieval of the vector size more generic ... This also enables the use of the function with "containers" that don't have a size function but implement an overload for `std::size()`, like c-arrays with fixed sizes.	2026-03-12 15:37:24 +01:00
Johannes Schmidt	32ca211ce5	Merge pull request #10681 from Icinga/api-console-missing-lock ... /v1/console: prevent concurrent use of the same session by multiple requests	2026-03-12 12:22:17 +01:00
Johannes Schmidt	2e2576c7c5	Add PerfdataWriterConnection class	2026-03-06 10:55:05 +01:00
Alexander Aleksandrovič Klimov	55d35b19dd	Merge pull request #10391 from Icinga/redis-query-variant ... Change Redis query type from `String` to an optional `std::string_view`	2026-03-04 10:38:12 +01:00
Johannes Schmidt	e58947dc45	Merge pull request #10737 from Icinga/better-debug-message-non-existing-host-service-dep ... Improve debug message when dependency child/parent don't exist	2026-03-04 10:36:12 +01:00
Johannes Schmidt	56c878d815	Improve debug message when dependency child/parent don't exist ... Previously this didn't print which of host/service is referenced and what the referenced non-existing checkable's name should have been. This commit adds that information.	2026-03-03 11:59:44 +01:00
Julian Brost	0d376b5d5a	/v1/console: prevent concurrent use of the same session by multiple requests ... If there are such requests, without this change, they would all be allowed and processed, resulting in unsafe concurrent (write) access to these data structures, which can ultimately crash the daemon or lead to other unintended behavior.	2026-03-03 11:32:39 +01:00
Julian Brost	0f44133786	Merge pull request #10732 from Icinga/icingadb-efficient-queue-helpers ... RedisConnection class enhancements & refactoring	2026-03-03 11:30:40 +01:00
Alexander A. Klimov	1702eebd41	Change RedisConnection::Query::value_type to a string/view variant ... Especially our history messages contain lots of hardcoded C string literals `"like this one"`. At runtime, they get translated to pointers to constant global memory, `const char*`. String `malloc(3)`s and copies these data every time. In contrast, the new type just stores the address if any. (Actually, `const char*` is wrapped by `std::string_view` to not compute its length every time.)	2026-03-03 11:29:07 +01:00
Julian Brost	851440a583	Merge pull request #10560 from Icinga/influxdb-mist ... Don't expect `Contet-Type` header to exactly match `application/json`	2026-03-03 11:28:45 +01:00
Yonas Habteab	c1763e4753	ObjectLock: allow to TryLock the mutex	2026-03-02 15:36:59 +01:00
Yonas Habteab	dd5289bf4e	RedisConnection: drop superfluous GetConnected method	2026-03-02 15:36:59 +01:00
Yonas Habteab	e0c239161b	Allow to track query stats in child Redis connections ... So that when we want the query stats of this specific connection we can easily get them, since the session leader contains the aggregated stats of all its children.	2026-03-02 15:36:58 +01:00
Yonas Habteab	4d4650b86a	RedisConnection: introduce & use RedisConnInfo helper struct	2026-03-02 15:34:04 +01:00
Alexander A. Klimov	c351f6a88c	Use RedisConnection::Quer* type aliases ... This expresses what kind of vector it is and allows to easily change those types in the future.	2026-03-02 12:45:09 +01:00
Alexander A. Klimov	ba1bb0e758	Introduce RedisConnection::QueryArg, an alias of String ... to ease changing that type in the future.	2026-03-02 12:45:09 +01:00
Alexander A. Klimov	808a6c77e8	IcingaDB#AddObjectDataToRuntimeUpdates(): take String by value ... It was copied anyway.	2026-03-02 12:45:09 +01:00
Johannes Schmidt	240142bcbd	Refactor OpenTsdbWriter to use a WorkQueue	2026-02-26 13:31:10 +01:00
Johannes Schmidt	7955d04ae3	Import std::chrono_literals into icinga namespace	2026-02-26 13:31:10 +01:00
Yonas Habteab	f934aa557b	Don't expect Contet-Type header to exactly match application/json ... InfluxDB seems to always include some other metadata in the `Content-Type` header separated by semicolons like `application/json; charset=utf-8`, and comparing the whole string with `application/json` will always fail, resulting in almost always useless warnings without any helpful information for the user what actually went wrong.	2026-02-25 15:55:55 +01:00
William Calliari	11726b741c	Take a mutex before accessing the l_ApiScriptFrames ... Take a mutex to avoid race conditions in the map that lead to segmentation faults. Move the ApiScriptFrame object back behind a shared pointer to avoid holding the mutex for too long. Fixes #10674	2026-02-25 08:37:03 +01:00
Johannes Schmidt	27c954dde8	Merge pull request #10715 from Icinga/fix-ubsan-errors ... Fix undefined behavior discovered with `-fsanitize=undefined`	2026-02-19 08:38:40 +01:00
Julian Brost	d02cdda5e9	Merge pull request #10716 from Icinga/drop-thread-local-variable-apiuser ... Remove `AuthenticatedApiUser` thread-local variable & pass it as arg instead	2026-02-13 14:43:36 +01:00
Johannes Schmidt	e976450bbb	Fix undefined behavior introduced by initializing last states with 99	2026-02-12 16:00:36 +01:00
Julian Brost	f5d5357fe2	Merge pull request #10028 from RincewindsHat/node_setup_no_globals ... Add cli option to disable the default global zones	2026-02-12 14:44:50 +01:00
Lorenz Kästle	32f5beaad3	Update lib/cli/nodesetupcommand.cpp ... Co-authored-by: Alexander Aleksandrovič Klimov <al2klimov@gmail.com>	2026-02-12 11:48:58 +01:00
Yonas Habteab	3b80153848	Remove AuthenticatedApiUser thread-local variable & pass it as param instead	2026-02-11 11:39:57 +01:00
Yonas Habteab	d4d46a9780	HTTP: stream responses where appropriate	2026-02-11 09:47:39 +01:00
Yonas Habteab	32887884e5	Make ValueGenerator more flexible & easy to use ... This commit refactors the ValueGenerator class to be a template that can work with any container type. Previously, one has to manually take care of the used container by lazily iterating over it within a lambda. Now, the `ValueGenerator` class itself takes care of all the iteration, making it easier to use and less error-prone. The new base `Generator` class is required to allow the `JsonEncoder` to handle generators in a type-erased manner.	2026-02-10 16:57:56 +01:00
Johannes Schmidt	216e2d9861	Fix uninitialized member in ScriptError	2026-02-06 12:10:20 +01:00
Alexander A. Klimov	56f8307c9a	ProcessSpawnImpl(): outsource array build and Convert::ToString() to child	2026-02-05 14:31:04 +01:00
Alexander A. Klimov	c5c572c4a2	Remove unused execvpe(3) copy	2026-02-05 14:31:01 +01:00
Alexander A. Klimov	f6124f39a3	ProcessSpawnImpl(): use POSIX execvp(3), not GNU/OpenBSD-only execvpe(3) ... Modify the fork(2)ed process' env in-place, then call execvp(3).	2026-02-05 14:30:42 +01:00
Yonas Habteab	91c7e60df8	Replace all existing copyright headers with SPDX headers ... I've used the following command to replace the original copyright header lines in a C-style comment block: ``` $ find . \( -type d \( -name '\..*' -o -name third-party -o -name scripts -o -name prefix -o -name malloc -o -name server -o -name docker -o -name build -o -name doc \) -prune \) -o -type f -exec perl -pi -e 's{/\*[^*]*\(\s*c\s*\)\s*(\d{4})\s*Icinga\s+GmbH[^*]*\*/}{// SPDX-FileCopyrightText: \1 Icinga GmbH <https://icinga.com>\n// SPDX-License-Identifier: GPL-2.0-or-later}gi' {} + ``` For files that use shell-style comments (#) like CMakeLists.txt, I've used this command: ``` $ find . \( -type d \( -name '\..*' -o -name third-party -o -name scripts -o -name prefix -o -name malloc -o -name server -o -name docker -o -name build -o -name doc \) -prune \) -o -type f -exec perl -pi -e 's{#.*\(\s*c\s*\)\s(\d{4})\sIcinga\s+GmbH.*}{# SPDX-FileCopyrightText: \1 Icinga GmbH <https://icinga.com>\n# SPDX-License-Identifier: GPL-2.0-or-later}gi' {} + ``` And for SQL files: ``` $ find . \( -type d \( -name '\..*' -o -name third-party -o -name scripts -o -name prefix -o -name malloc -o -name server -o -name docker -o -name build -o -name doc \) -prune \) -o -type f \( -name '*.sql' \) -exec perl -pi -e 's{--.*\(c\)\s(\d{4})\sIcinga\sGmbH.*}{-- SPDX-FileCopyrightText: \1 Icinga GmbH <https://icinga.com>\n-- SPDX-License-Identifier: GPL-2.0-or-later}gi' {} + $ find . \( -type d \( -name '\..*' -o -name third-party -o -name scripts -o -name prefix -o -name malloc -o -name server -o -name docker -o -name build -o -name doc \) -prune \) -o -type f \( -name '*.sql' \) -exec perl -pi -e 's{-- Copyright \(c\)\s(\d{4})\sIcinga\s+Development\sTeam.*}{-- SPDX-FileCopyrightText: \1 Icinga GmbH <https://icinga.com>\n-- SPDX-License-Identifier: GPL-2.0-or-later}gi' {} + ```	2026-02-04 14:00:05 +01:00
Yonas Habteab	36b41dbe11	Merge pull request #10698 from ymartin-ovh/pr-10687 ... Fix nullptr dereference on `{child_,parent_}service` navigation	2026-01-30 16:49:19 +01:00
Alexander Aleksandrovič Klimov	9bffe06169	Merge pull request #10388 from Icinga/Registry-Freeze ... Freeze registries at startup, when everything has been registered	2026-01-26 17:28:43 +01:00
Alexander A. Klimov	6223bd5f31	Remove unused DbTypeRegistry	2026-01-26 14:37:59 +01:00
Alexander A. Klimov	b4192bd80a	Replace class B : public A { }; with using B = A; (refactor only)	2026-01-26 14:34:29 +01:00
Alexander A. Klimov	74ac0183ca	Registry<U,T>: remove unused template typename U	2026-01-26 10:24:22 +01:00
Alexander Aleksandrovič Klimov	363a051e18	Merge pull request #9728 from Icinga/silence-compiler-warnings-about-unused-parameters ... Silence compiler warnings about unused parameters	2026-01-23 16:40:54 +01:00
Alexander A. Klimov	4ba46f9eb2	Silence compiler warnings about unused parameters ... Every of these parameters exists for a reason. The best we can do is to convince the compiler.	2026-01-23 13:31:01 +01:00
Alexander Aleksandrovič Klimov	92d9f64eb0	Merge pull request #10693 from Icinga/AddDowntime-origin ... `Downtime::AddDowntime()`: remove unused parameter `const MessageOrigin::Ptr& origin`	2026-01-23 13:22:55 +01:00
Alexander A. Klimov	92b6034504	Downtime::AddDowntime(): remove unused parameter const MessageOrigin::Ptr& origin	2026-01-23 10:41:18 +01:00
Johannes Schmidt	1505f09ed6	Refactor HttpMessage into generalized templated types ... This adds generalized IncomingHttpMessage and OutgoingHttpMessage templates that support different types of streams (via a std::variant) and can both be used for either requests or responses. The tacked on metadata from the old HttpRequest and server connection from the old HttpServerConnection have been moved to HttpApi(Request|Response) classes that derive from the above generalized message types.	2026-01-22 17:20:32 +01:00
Johannes Schmidt	a0f603f608	Update names of HttpRequest and HttpResponse	2026-01-22 12:41:21 +01:00
Yannick Martin	1f7c38c792	Fix(#10687): address null pointer dereference on navigate	2026-01-16 18:14:46 +01:00
Johannes Schmidt	f493f2bbca	Fix a race condition in NotificationComponent ... Fix the race condition described in the previous commit by locking the checkable earlier in NotificationTimerHandler().	2026-01-13 10:55:29 +01:00
Egor-OSSRevival	0d32ae3159	docs: Remove 'queue' parameter requirement from event stream document… (#10495) ... * docs: Remove 'queue' parameter requirement from event stream documentation * Update AUTHORS	2026-01-07 14:51:26 +01:00
Julian Brost	dc09713ac4	Merge pull request #10350 from Icinga/unittest-certificate-verification ... Test internal cert generation & verification process	2026-01-07 12:00:28 +01:00
Yonas Habteab	0b2f22ce40	tlsutility: make cert ts configurable & use ASN1_TIME_compare for comparison	2025-12-04 13:01:48 +01:00
Julian Brost	fa3063d148	IsValidHeaderValue: use front()/back() instead of iterators ... Don't ask me why I wasn't thinking of the very basic front() and back() methods when writing this code. Does exactly the same here, but is much more straight-forward than the extra iterator detour.	2025-12-03 11:09:44 +01:00
Julian Brost	985db970bb	Allow to set extra headers in HTTP responses ... Use case: Allow settings headers like Strict-Transport-Security if one likes. How this headers would benefit the Icinga 2 API is questionable, but there are security scanners that see HTTPS and complain about it, so this gives an easy way to make them happy (with this probably being the only benefit).	2025-11-28 16:19:19 +01:00
Julian Brost	aca67f6d88	HttpUtility: add functions for validating HTTP header names and values	2025-11-28 16:19:19 +01:00
Brad Smith	bbc1859aa1	Have thread name length clamping also apply for the other OSes ... This covers macOS, *BSDs, Linux, Cygwin. NetBSD PTHREAD_MAX_NAMELEN_NP 32 macOS 64 OpenBSD 24 Linux/FreeBSD/DragonFly/Cygwin 16	2025-11-27 09:19:16 -05:00
Brad Smith	7aa777c40f	Fix usage of pthread_setname_np() on NetBSD ... NetBSD's pthread_setname_np() takes a slightly different set of parameters.	2025-11-27 07:16:34 -05:00
Alvar	8c230df3da	Merge pull request #10640 from brad0/openbsd_dlopen ... Fix dlopen()'ing libraries on OpenBSD	2025-11-27 12:13:18 +00:00
Brad Smith	d243da72e5	Fix dlopen()'ing libraries on OpenBSD	2025-11-27 04:52:28 -05:00
Brad Smith	cf8fc3d64f	Add tests for pthread_set_name_np() and pthread_setname_np() ... Looks like the conversion from autoconf to CMake dropped these tests along the way.	2025-11-27 10:48:54 +01:00
Brad Smith	77881a7e18	Enable the pthread_np.h header on OpenBSD and DragonFly ... Necessary for pthread_set_name_np().	2025-11-25 15:46:02 -05:00
Yonas Habteab	ed9014103f	Merge pull request #10609 from Icinga/fix-misc-compiler-warnings ... Fix misc compiler warnings	2025-11-19 14:26:57 +01:00
Johannes Schmidt	11e83b6274	Fix -Wunused-lambda-capture warning	2025-11-19 12:00:58 +01:00
Johannes Schmidt	a2e0ce426f	Fix -Wunused-variable warnings	2025-11-19 12:00:58 +01:00
Johannes Schmidt	8ef1604f72	Fix -Wunused-but-set-variable warnings (clang)	2025-11-19 12:00:58 +01:00
Johannes Schmidt	6bd89821cd	Fix -Wenum-compare warnings	2025-11-19 12:00:58 +01:00
Johannes Schmidt	7b4697a237	Fix -Wredundant-move warnings	2025-11-19 12:00:58 +01:00
Johannes Schmidt	93e69ed0e8	Fix -Wsign-compare warnings	2025-11-19 12:00:58 +01:00
Johannes Schmidt	758d1b6873	Fix -Wsuggest-override warnings	2025-11-19 12:00:58 +01:00
Johannes Schmidt	4f0a24f8ab	Add ABORT() macro for unconditionally failing ASSERTS ... This is necessary to stop MSVC complaining that "not all paths return a value", because it is not able to infer that the expression `false ? 0 : non_returning_function()` never returns. In the process of debugging this I've also slightly simplified the other assert macros and abort function, so they don't need compiler specific preprocessor-paths anymore.	2025-11-19 11:55:49 +01:00
Johannes Schmidt	9a69dd9a45	Use 64bit unsigned integer for Replay log timestamps	2025-11-19 09:41:13 +01:00