cysystemd 1.6.0

Logging handler for python logger

from cysystemd import journal
import logging
import uuid

logging.basicConfig(level=logging.DEBUG)
logger = logging.getLogger()
logger.addHandler(journal.JournaldLogHandler())

try:
    logger.info("Trying to do something")
    raise Exception('foo')
except:
    logger.exception("Test Exception %s", 1)

systemd daemon notification

from cysystemd.daemon import notify, Notification

# Send READY=1
notify(Notification.READY)

# Send status
notify(Notification.STATUS, "I'm fine.")

# Send stopping
notify(Notification.STOPPING)

Write message into systemd journal

from cysystemd import journal


journal.write("Hello Lennart")

# Or send structured data
journal.send(
    message="Hello Lennart",
    priority=journal.Priority.INFO,
    some_field='some value',
)

Reading all systemd records

from cysystemd.reader import JournalReader, JournalOpenMode

journal_reader = JournalReader()
journal_reader.open(JournalOpenMode.SYSTEM)
journal_reader.seek_head()

for record in journal_reader:
   print(record.data['MESSAGE'])

Read only cron logs

from cysystemd.reader import JournalReader, JournalOpenMode, Rule


rules = (
   Rule("SYSLOG_IDENTIFIER", "CRON") &
   Rule("_SYSTEMD_UNIT", "crond.service") |
   Rule("_SYSTEMD_UNIT", "cron.service")
)

cron_reader = JournalReader()
cron_reader.open(JournalOpenMode.SYSTEM)
cron_reader.seek_head()
cron_reader.add_filter(rules)

for record in cron_reader:
   print(record.data['MESSAGE'])

Polling records

from cysystemd.reader import JournalReader, JournalOpenMode


reader = JournalReader()
reader.open(JournalOpenMode.SYSTEM)
reader.seek_tail()

poll_timeout = 255

while True:
   reader.wait(poll_timeout)

   for record in reader:
      print(record.data['MESSAGE'])

journald open modes

• CURRENT_USER

• LOCAL_ONLY

• RUNTIME_ONLY

• SYSTEM

• SYSTEM_ONLY

from cysystemd.reader import JournalReader, JournalOpenMode

reader = JournalReader()
reader.open(JournalOpenMode.CURRENT_USER)

journald entry

JournalEntry class has some special properties and methods:

• data - journal entry content (dict)

• date - entry timestamp (datetime instance)

• cursor - systemd identification bytes for this entry

• boot_id() - returns bootid

• get_realtime_sec() - entry epoch (float)

• get_realtime_usec() - entry epoch (int microseconds)

• get_monotonic_sec() - entry monotonic time (float)

• get_monotonic_usec() - entry monotonic time (int microseconds)

• __getitem__(key) - shoutcut for entry.data[key]

journald reader

JournalReader class has some special properties and methods:

• open(flags=JournalOpenMode.CURRENT_USER) - opening journald with selected mode

• open_directory(path) - opening journald from path

• open_files(*filename) - opening journald from files

• data_threshold - may be used to get or set the data field size threshold for data returned by fething entry data.

• closed - returns True when journal reader closed

• locked - returns True when journal reader locked

• idle - returns True when journal reader opened

• seek_head - move reader pointer to the first entry

• seek_tail - move reader pointer to the last entry

• seek_monotonic_usec - seeks to the entry with the specified monotonic timestamp, i.e. CLOCK_MONOTONIC. Since monotonic time restarts on every reboot a boot ID needs to be specified as well.

• seek_realtime_usec - seeks to the entry with the specified realtime (wallclock) timestamp, i.e. CLOCK_REALTIME. Note that the realtime clock is not necessarily monotonic. If a realtime timestamp is ambiguous, it is not defined which position is sought to.

• seek_cursor - seeks to the entry located at the specified cursor (see JournalEntry.cursor).

• wait(timeout) - It will synchronously wait until the journal gets changed. The maximum time this call sleeps may be controlled with the timeout_usec parameter.

• __iter__ - returns JournalReader object

• __next__ - calls next() or raise StopIteration

• next(skip=0) - returns the next JournalEntry. The skip parameter skips some entries.

• previous(skip=0) - returns the previous JournalEntry. The skip parameter skips some entries.

• skip_next(skip) - skips next entries.

• skip_previous(skip) - skips next entries.

• add_filter(rule) - adding filter rule. See read-only-cron-logs as example.

• clear_filter - reset all filters

• fd - returns a special file descriptor

• events - returns EPOLL events

• timeout - returns internal timeout

• process_events() - After each poll() wake-up process_events() needs to be called to process events. This call will also indicate what kind of change has been detected.

• get_catalog() - retrieves a message catalog entry for the current journal entry. This will look up an entry in the message catalog by using the “MESSAGE_ID=” field of the current journal entry. Before returning the entry all journal field names in the catalog entry text enclosed in “@” will be replaced by the respective field values of the current entry. If a field name referenced in the message catalog entry does not exist, in the current journal entry, the “@” will be removed, but the field name otherwise left untouched.

• get_catalog_for_message_id(message_id: UUID) - works similar to get_catalog() but the entry is looked up by the specified message ID (no open journal context is necessary for this), and no field substitution is performed.

AsyncJournalReader

Blocking methods were wrapped by threads. Method wait() use epoll on journald file descriptor.

import asyncio
import json

from cysystemd.reader import JournalOpenMode
from cysystemd.async_reader import AsyncJournalReader


async def main():
    reader = AsyncJournalReader()
    await reader.open(JournalOpenMode.SYSTEM)
    await reader.seek_tail()

    while await reader.wait():
        async for record in reader:
            print(
                json.dumps(
                    record.data,
                    indent=1,
                    sort_keys=True
                )
            )

if __name__ == '__main__':
    asyncio.run(main())