aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
...
| * mod_c2s, mod_s2s, mod_component, mod_bosh, mod_websockets: Set default ↵Matthew Wild2021-05-075-5/+7
| | | | | | | | | | | | | | | | | | | | | | | | stanza size limits c2s/bosh/ws streams will default to 256KB, s2s and components to 512KB. These values are aligned with ejabberd's default settings, which should reduce issues related to inconsistent size limits between servers on the XMPP network. The previous default (10MB) is excessive for any production server, and allows significant memory usage by even unauthenticated sessions.
| * util.startup: Set more aggressive defaults for GCMatthew Wild2021-05-071-1/+7
| | | | | | | | | | | | | | | | | | Testing has demonstrated that the default GC parameters are not sufficient to prevent runaway memory growth when running under Lua 5.2 and Lua 5.3. Setting the GC speed to 500 was tested on Lua versions 5.1->5.4 and did not display unbounded memory growth.
| * prosodyctl: check config: Add 'gc' to list of global optionsMatthew Wild2021-04-301-1/+1
| |
* | core.certmanager: Catch error from lfsKim Alvefur2021-05-071-1/+11
| | | | | | | | | | | | | | lfs.dir() throws a hard error if there's a problem, e.g. no such directory or permission issues. This also gets called early enough that the main loop error protection hasn't been brought up yet, causing a proper crash.
* | core.certmanager: Resolve certs path relative to config dirKim Alvefur2021-05-071-3/+3
| | | | | | | | | | | | | | | | | | Otherwise the default "certs" would be relative to $PWD, which works when testing from a source checkout, but not on installed systems where it usually points to the data directory. Also, the LuaFileSystem dir() iterator throws a hard error, which may cause a crash or other problems.
* | core.portmanager: Use existing http_host for https SNI mappingKim Alvefur2021-05-051-0/+4
| |
* | core.portmanager: Allow overriding expected SNI name per serviceKim Alvefur2021-05-051-2/+3
| | | | | | | | | | | | | | E.g. VirtualHost"example.com" https_name = "xmpp.example.com"
* | core.certmanager: Skip directly to guessing of key from cert filenameKim Alvefur2021-05-051-2/+17
| | | | | | | | Cuts down on a ton of debug logs
* | core.certmanager: Join paths with OS-aware util.paths functionKim Alvefur2021-05-051-2/+3
| | | | | | | | Right thing to do, rather than hardcoding '/'
* | core.certmanager: Build an index over certificatesKim Alvefur2021-04-101-0/+79
| |
* | core.certmanager: Check for complete filenameKim Alvefur2021-04-101-1/+1
| | | | | | | | Prevents a false positive match on files with fullchain.pem as suffix
* | mod_tls: Fix order of debug messages and tls context creationKim Alvefur2021-05-051-2/+2
| | | | | | | | | | Originally added in 5b048ccd106f Merged wrong in ca01c449357f
* | configure: Make ICU the default library for IDNA and JID normalizationKim Alvefur2020-07-121-3/+3
| | | | | | | | See #533
* | net.http.server: Split out method for sending only the headerKim Alvefur2021-04-241-2/+7
| | | | | | | | | | Makes it easier to reuse, e.g. for SSE or websockets or other custom responses.
* | s2s et al.: Add counters for connection state transitionsJonas Schäfer2021-04-213-11/+50
| |
* | mod_s2s: Port to new OpenMetrics APIJonas Schäfer2021-04-181-10/+26
| |
* | mod_c2s: Port to new OpenMetrics APIJonas Schäfer2021-04-181-10/+16
| |
* | Statistics: Rewrite statistics backends to use OpenMetricsJonas Schäfer2021-04-187-494/+1110
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The metric subsystem of Prosody has had some shortcomings from the perspective of the current state-of-the-art in metric observability. The OpenMetrics standard [0] is a formalization of the data model (and serialization format) of the well-known and widely-used Prometheus [1] software stack. The previous stats subsystem of Prosody did not map well to that format (see e.g. [2] and [3]); the key reason is that it was trying to do too much math on its own ([2]) while lacking first-class support for "families" of metrics ([3]) and structured metric metadata (despite the `extra` argument to metrics, there was no standard way of representing common things like "tags" or "labels"). Even though OpenMetrics has grown from the Prometheus world of monitoring, it maps well to other popular monitoring stacks such as: - InfluxDB (labels can be mapped to tags and fields as necessary) - Carbon/Graphite (labels can be attached to the metric name with dot-separation) - StatsD (see graphite when assuming that graphite is used as backend, which is the default) The util.statsd module has been ported to use the OpenMetrics model as a proof of concept. An implementation which exposes the util.statistics backend data as Prometheus metrics is ready for publishing in prosody-modules (most likely as mod_openmetrics_prometheus to avoid breaking existing 0.11 deployments). At the same time, the previous measure()-based API had one major advantage: It is really simple and easy to use without requiring lots of knowledge about OpenMetrics or similar concepts. For that reason as well as compatibility with existing code, it is preserved and may even be extended in the future. However, code relying on the `stats-updated` event as well as `get_stats` from `statsmanager` will break because the data model has changed completely; in case of `stats-updated`, the code will simply not run (as the event was renamed in order to avoid conflicts); the `get_stats` function has been removed completely (so it will cause a traceback when it is attempted to be used). Note that the measure_*_event methods have been removed from the module API. I was unable to find any uses or documentation and thus deemed they should not be ported. Re-implementation is possible when necessary. [0]: https://openmetrics.io/ [1]: https://prometheus.io/ [2]: #959 [3]: #960
* | util.prosodyctl.shell: Fix for different location of unpack in Lua 5.1Kim Alvefur2021-04-251-1/+2
| |
* | util.prosodyctl.shell: Fix for missing 'sep' arg to string.rep in Lua 5.1Kim Alvefur2021-04-251-1/+2
| |
* | util.prosodyctl.shell: Allow calling console commands with fewer shell quotesKim Alvefur2021-04-251-3/+2
| | | | | | | | | | | | | | | | E.g. `prosodyctl shell module reload disco example.com` becomes equivalent to `prosodyctl shell 'module:reload("disco", "example.com")`. Won't work for every possible command, but reduces the amount of shell quoting problems for most common commands.
* | mod_tls: Bail out if session got destroyed while sending <proceed/>Kim Alvefur2021-04-151-0/+1
| | | | | | | | | | | | | | Can happen in case opportunistic_writes is enabled and the session got destroyed while writing that tag. Thanks Ge0rG
* | core.sessionmanager: Add a dummy rawsend() method to resting sessionsKim Alvefur2021-04-151-0/+1
| | | | | | | | | | | | | | Should fix a traceback on attempted use after destruction, in case where opportunistic_writes was in use. Thanks Ge0rG
* | mod_c2s: Fix traceback if session was destroyed while opening stream (thanks ↵Kim Alvefur2021-04-141-0/+3
| | | | | | | | | | | | | | | | | | | | | | Ge0rG) Could happen with the 'opportunistic_writes' setting, since then the stream opening is written directly to the socket, which can in turn trigger session destruction if the socket somehow got closed just after the other sent their stream header. Error happens later when it tries to `hosts[session.host == nil].events`
* | util.xmppstream: Send <?xml?> declaration in same call as stream openKim Alvefur2021-04-141-2/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | If network_settings.opportunistic_writes is enabled then this would previously have resulted in two socket writes, and possibly two packets being sent. This caused some issues in older versions of Gajim, which apparently expected the stream opening in the first packet, and thus it could not connect. With this change and opportunistic_writes enabled, the first packet should contain both the xml declaration and the stream open tag. Without opportunistic_writes, there should be no observable change. Tested with Gajim 1.1.2 (on same machine). Unsure if loopback behaves differently than the network here.
* | core.statsmanager: Allow special "manual" value for statistics_intervalKim Alvefur2021-04-062-6/+19
| | | | | | | | | | | | | | | | | | | | | | | | | | When set, no periodic statistics collection is done by core.statsmanager, instead some module is expected to call collect() when it suits. Obviously only one such module should be enabled. Quoth jonas’ > correct way is to scrape the internal sources on each call to /metrics > in the context of Prometheus "manual" as opposed to "automatic", from the point of view of statsmanager.
* | Merge 0.11->trunkKim Alvefur2021-04-102-4/+15
|\|
| * mod_saslauth: Use a defined SASL errorKim Alvefur2021-03-181-1/+1
| |
| * mod_saslauth: Improve code styleKim Alvefur2021-03-181-1/+3
| | | | | | | | | | | | This many returns deserve their own line. `session["sasl_handler"]` style isn't used anywhere else.
| * mod_c2s: Log about missing conn on async state changesKim Alvefur2021-03-181-0/+4
| |
| * mod_c2s: Improve code styleKim Alvefur2021-03-181-2/+6
| | | | | | | | | | We don't use the quoted table indexing style that often, it's not needed here and it's enough to check for falsyness rather than `nil`.
| * mod_c2s: Fix traceback in session close when conn is nilKim Alvefur2021-03-171-2/+2
| | | | | | | | Unclear how this happens.
| * mod_saslauth: Don't throw errors in async code when connections are gonetmolitor2021-03-181-0/+1
| | | | | | | | Fixes #1515
| * mod_c2s: Don't throw errors in async code when connections are gonetmolitor2021-03-181-2/+2
| | | | | | | | Fixes #1507
* | core.statsmanager: Cover util.statistics work in processing measurementKim Alvefur2021-04-061-1/+1
| | | | | | | | | | | | | | | | | | Zash> Btw, this conditional and loop, shouldn't it be covered by the timing measurement? Zash> Isn't that where all the util.statistics work is done? MattJ> Yeah, it should Zash> ("the", but there's two ... which one‽) MattJ> Yeah... not sure :) MattJ> Processing I guess
* | core.hostmanager: Remove obsolete clearing of srv_hostsKim Alvefur2021-04-061-1/+0
| | | | | | | | Gone with s2sout.lib in 756b8821007a
* | mod_admin_shell: Remove obsolete checks related to s2sout.libKim Alvefur2021-04-061-14/+1
| | | | | | | | | | s2sout.lib was removed in 756b8821007a along with srv_hosts and srv_choice
* | mod_http_file_share: Include expiry time of the upload itself in tokenKim Alvefur2021-04-051-0/+1
| | | | | | | | | | | | Lets an external upload service know this so it can do expiry itself. Could possibly have been calculated based on the token expiry or issuance time, explicit > implicit.
* | mod_http_file_share: Include time of issuance in auth tokenKim Alvefur2021-04-051-1/+3
| | | | | | | | | | In case an external upload service wants to have the original creation time, or calculate the token expiry itself.
* | mod_http_file_share: Group related properties for readabilityKim Alvefur2021-04-051-2/+6
| |
* | mod_http_file_share: Log error opening file for writingKim Alvefur2021-04-051-2/+3
| | | | | | | | util.error.coerce() doesn't work well with iolib
* | mod_http_file_share: Fix logging of error opening fileKim Alvefur2021-04-051-1/+1
| | | | | | | | It's annoying that Lua interpolates the filename into the error message.
* | MUC: Allow overriding occupant object from groupchat message eventKim Alvefur2021-03-251-4/+9
| | | | | | | | | | | | | | Use case: Enable module that provides a virtual occupant object for bots Before, if there is no occupant then either some other part of MUC would reject the message or `occupant.nick` would have caused an error.
* | net.server_epoll: Remove unnecessary luacheck annotationsKim Alvefur2021-04-011-3/+3
| | | | | | | | | | Not sure why these were here to begin with, since it does use the 'self' argument and did so since they were added.
* | mod_http_file_share: Include storage in prune time measurementKim Alvefur2021-03-311-1/+1
| | | | | | | | Maybe the original idea was that you would measure storage separately?
* | mod_http_file_share: Add internal command to check files consistencyKim Alvefur2021-03-281-0/+18
| | | | | | | | | | | | Background: Found a few files in my store that did not match the size recorded in the slot, so I needed a way to check which which those were. As it was a bit too much to type into the shell I added it here instead.
* | util.datamapper: Fix spelling in testsKim Alvefur2021-03-281-2/+2
| |
* | mod_http_file_share: Fix reporting of missing filesKim Alvefur2021-03-281-2/+5
| | | | | | | | | | | | | | This just gave an unhelpful 500 error. It would be nice to have some wrapper code that could untangle the embedded filename in the io libs errors.
* | MUC: Report number of live rooms to statsmanagerKim Alvefur2021-03-281-0/+5
| | | | | | | | Should help inform on whether the cache size should be increased.
* | mod_http_file_share: Report number of items in caches to statsmanagerKim Alvefur2021-03-281-0/+8
| | | | | | | | | | | | This is neat, O(1) reporting, why don't we do this everywhere? Gives you an idea of how much stuff is in the caches, which may help inform decisions on whether the size is appropriate.