Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
最近の監視(仮)/Recent system monitoring with mackerel
Search
kazeburo
December 23, 2019
Technology
5.4k
3
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
最近の監視(仮)/Recent system monitoring with mackerel
mackerel day #2
kazeburo
December 23, 2019
More Decks by kazeburo
See All by kazeburo
さくらのクラウド開発ふりかえり2025
kazeburo
2
3.4k
国産クラウドを支える設計とチームの変遷 “技術・組織・ミッション”
kazeburo
7
18k
クラウド開発の舞台裏とSRE文化の醸成 / SRE NEXT 2025 Lunch Session
kazeburo
1
2.6k
さくらのクラウド 開発の挑戦とその舞台裏
kazeburo
1
1.5k
[SRE kaigi 2025] ガバメントクラウドに向けた開発と変化するSRE組織のあり方 / Development for Government Cloud and the Evolving Role of SRE Teams
kazeburo
4
4.4k
[さくらのTech Day] ガバメントクラウド開発と変化と成長する組織 / sakura techday, Develop govcloud and the team
kazeburo
0
9.4k
ガバメントクラウド開発と変化と成長する組織 / Organizational change and growth in developing a government cloud
kazeburo
4
3.8k
DNS水責め攻撃と監視 / DNS water torture attack Monitoring and SLO
kazeburo
5
4.6k
DBやめてみた / DNS water torture attack and countermeasures
kazeburo
13
14k
Other Decks in Technology
See All in Technology
10年目を迎えた「ABEMA」がどのように AI 活用を推進して、AI 駆動開発にシフトしているのか / How ABEMA, entering its 10th year, is promoting the use of AI and shifting toward AI-driven development
miyukki
0
130
Control Planeで育てるBtoB SaaSの認証基盤 - SRE NEXT 2026
pokohide
1
2.3k
技術イベント終了後、運営の 事後タスクは丁寧に (心がけています)/ #tamagawadev
nishiuma
1
100
「最後に責任を取るのはチーム」— 人間のPRレビューを最小化してアップデートしたメンタルモデル
jnishime_dresscode
0
600
ポストモーテム! DDoSからサイトは守れた。 でもビジネスは守れなかった。
bengo4com
0
2.9k
脱金融のフューチャー・デザイン / Future Design Beyond Finance
ks91
PRO
0
150
しぶいSRE: サーバから見えない障害にどう向き合うか。ラストワンマイルのデバッグ実践 / Shibui SRE
kanny
13
6.1k
Keeping applications secure by evolving OAuth 2.0 and OpenID Connect
ahus1
PRO
1
160
事業価値を⽣み出すSREへ SREが担うべき意思決定の5層
kenta_hi
2
3.6k
Mastraエージェント、どのクラウドにデプロイする?
minorun365
PRO
2
180
世界、断片、モデル。そして理解
ardbeg1958
1
110
アカウントが増えてからでは遅い? ~ マルチアカウント統制の勘所 ~
kenichinakamura
0
230
Featured
See All Featured
Game over? The fight for quality and originality in the time of robots
wayneb77
1
220
Why Your Marketing Sucks and What You Can Do About It - Sophie Logan
marketingsoph
0
230
Leveraging Curiosity to Care for An Aging Population
cassininazir
1
340
So, you think you're a good person
axbom
PRO
2
2.1k
Exploring the relationship between traditional SERPs and Gen AI search
raygrieselhuber
PRO
2
4.1k
The World Runs on Bad Software
bkeepers
PRO
72
12k
The Straight Up "How To Draw Better" Workshop
denniskardys
239
140k
Odyssey Design
rkendrick25
PRO
2
730
Technical Leadership for Architectural Decision Making
baasie
3
440
Building the Perfect Custom Keyboard
takai
2
810
JAMstack: Web Apps at Ludicrous Speed - All Things Open 2022
reverentgeek
1
490
Design in an AI World
tapps
1
260
Transcript
࠷ۙͷࢹ(Ծ) Mackerel Day #2 kazeburo
Me • Masahiro Nagano • @kazeburo • Mackerel Ξϯόαμʔ •
Ұ൪͖ͳ mackerelϓϥάΠϯ mackerel-plugin-axslog (ࣗ࡞)ͩͳɻ ʹཱͭͱ͜ΖͰ diff-detector(ࣗ࡞)
https://ihara2525.tumblr.com/ “Ұߦͷϩάͷ͜͏ʹɺҰਓͷϢʔβ͕͍Δ”
• ΤϥʔόδΣοτͱ͍͏ߟ͑ • ೦ʹࢥ͏ϢʔβΛগͳ͍͖͍ͯͨ͘͠ • ࢹϢʔβͷظʹԠ͑ΔͨΊͷπʔϧ
• ʮࢹʯʹΑΓ • োΛະવʹ͙ • োʹૉૣ͘ؾͮ͘ • োͷݪҼಛఆͷώϯτ/MTTRͷॖ
mackerel-plugin-axslog ͱੜ
None
• mackerel-plugin-axslog • https://github.com/kazeburo/mackerel-plugin-axslog • mackerel-plugin-accesslog ͱಉ͘͡accesslogͷՄࢹԽ • status͝ͱͷͱׂ߹ɺresponse timeͷpercentileΛܭଌ
• ूܭ͢ΔlabelΛมߋՄೳɾJSONαϙʔτ • ߴ
0 2 4 6 8 accesslog(ltsv) axslog(ltsv) axslog(json) 1.17s 0.584s
7.976s 120ສߦ(20000req/sec)ͷϩάΛॲཧ͢Δ࣌ؒ
• ߴԽͷར • γεςϜෛՙΛؾʹͤͣɺʮΨϯΨϯʯϝτϦΫεΛ૿ͤΔ • ϑΟϧλػೳͷՃ
[plugin.metrics.axslog-web-googlebot] command = "/usr/local/bin/mackerel-plugin-axslog --key-prefix=googlebot \ --logfile /var/log/nginx/access.log --filter Googlebot"
[plugin.metrics.axslog-web-adsbot] command = "/usr/local/bin/mackerel-plugin-axslog --key-prefix=adsbot \ --logfile /var/log/nginx/access.log --filter AdsBot-Google"
• mackerel-plugin-postfix-log • https://github.com/kazeburo/mackerel-plugin-postfix-log • SMTP • AxlogͰlogrotateΛै͢Δίʔυॻ͍ͨͷͰ͙͢Ͱ͖ͨ
None
postfixlog.total_delay.average 0.240476 1555681849 postfixlog.total_delay.99_percentile 0.250000 1555681849 postfixlog.total_delay.95_percentile 0.240000 1555681849 postfixlog.total_delay.90_percentile
0.240000 1555681849 postfixlog.recving_delay.average 0.040000 1555681849 postfixlog.recving_delay.99_percentile 0.040000 1555681849 postfixlog.recving_delay.95_percentile 0.040000 1555681849 postfixlog.recving_delay.90_percentile 0.040000 1555681849 postfixlog.queuing_delay.average 0.000476 1555681849 postfixlog.queuing_delay.99_percentile 0.010000 1555681849 postfixlog.queuing_delay.95_percentile 0.000000 1555681849 postfixlog.queuing_delay.90_percentile 0.000000 1555681849 postfixlog.connection_delay.average 0.090000 1555681849 postfixlog.connection_delay.99_percentile 0.090000 1555681849 postfixlog.connection_delay.95_percentile 0.090000 1555681849 postfixlog.connection_delay.90_percentile 0.090000 1555681849 postfixlog.transmission_delay.average 0.090000 1555681849 postfixlog.transmission_delay.99_percentile 0.090000 1555681849 postfixlog.transmission_delay.95_percentile 0.090000 1555681849 postfixlog.transmission_delay.90_percentile 0.090000 1555681849 postfixlog.transfer_num.2xx_count 1.615385 1555681849 postfixlog.transfer_num.4xx_count 0.000000 1555681849 postfixlog.transfer_num.5xx_count 0.000000 1555681849 postfixlog.transfer_total.count 1.615385 1555681849 postfixlog.transfer_ratio.2xx_percentage 100.000000 1555681849 postfixlog.transfer_ratio.4xx_percentage 0.000000 1555681849 postfixlog.transfer_ratio.5xx_percentage 0.000000 1555681849 ड৴࣌ͷԆ Ωϡʔʹೖ͍ͬͯΔ࣌ؒ ૹ৴ઌͷଓʹ͔͔Δ࣌ؒ ૹ৴ͷԆ
• mackerel-plugin-log-incr-rate • https://github.com/kazeburo/mackerel-plugin-log-incr-rate • ͋Δϩάʹର͢Δผͷϩάͷߦͷ૿Ճׂ߹Λࢹ͢Δ • accesslogʹର͢Δapplication_log, errorlogͷ݅ͳͲ •
application_logͷॻ͖ࠐΈ͕രൃͯ͠ϩάܭଌʹӨڹͨ͠োʹର͢Δࢹ
None
• mackerel-plugin-gzip-response • Not OSS • ϨεϙϯεͷGzipѹॖ͕ߦΘΕ͍ͯΔ͔ࢹ • Ұ෦ΫϥΠΞϯτͷ࣮ͰGzip͕ແޮʹͳͬͨ··ɺେ͖ͳ৴ͱͳΔϚε λʔߋ৽͕ߦΘΕɺωοτϫʔΫʹෛՙ͕͔͔Δࣄଶ͕͋ͬͨ
• ϩάʹϨεϙϯεͷ Conent-Encoding ϔομΛه͠ूܭ
None
• Mackerel Plugin ࡞ύλʔϯ • ো/োʹܨ͕Δͷൃݟ • Metrics Plugin ͷ࡞ͱαʔόͷө
• MetricsΛΈͯͷɺᮢઃఆɾਵ࣌ݟ͠
DB ӡ༻ͱࢹ
• ϗετͷεςʔλεΛࢹ͢Δ • ϗετεςʔλε Standby ͷDB(slave)αʔόΛ production ʹೖ • ϝϯςφϯε࡞ۀͷͨΊϨϓϦέʔγϣϯ͕ࢭ·͍͕ͬͯͨɺ௨དྷͳ͍
App App consul consul DNS consul consul 1) look up
DNS 2) connect DB check mysql uptime before slave slave slave
App App consul consul DNS consul consul 1) look up
DNS 2) connect DB A) check mysql uptime After Mackerel batch batch batch B) check host status file B’ ) update host status file by cron slave slave slave
• Consul ͷαʔϏε͕༗ޮʹͳΔͨΊʹɺmackerelͷεςʔλεͷมߋ͕ ඞཁɻΞϥʔτʹؾ͕ͭ͘Մೳੑˢ • ϗετεςʔλεΛϩʔΧϧͷͲ͔͜ʹॻ͖ࠐΉػೳɺ mackerel-agent ʹ͋Δͱศར?
• Slave Ͱͷ Trigger ࣮ߦΛकΔ • ProductionͷҰ෦ͷslaveΛɺௐࠪɾूܭ༻ͱͯ͑͠ΔΑ͏ʹ͍ͯ͠Δ • ηϯγςΟϒͳใʹ͍ͭͯSlave্ͷTriggerͰআ
• MySQLͱϨϓϦέʔγϣϯͱTrigger • MySQLͷϨϓϦέʔγϣϯͷํࣜʹSQL͕సૹ͞Εͯ͘ΔStatementํࣜͱɺ σʔλ͕ૹΒΕͯ͘ΔRow Basedํࣜɺ·ͨͲͪΒ͔͕ࣗಈͰબ͞ΕΔMixed ํ͕ࣜ͋ΔɻMixedํࣜͰجຊStatementํ͕ࣜͩɺඇܾఆੑΫΤϦ͕Θ ΕΔͱɺRow BasedͱͳΔ •
ϨϓϦέʔγϣϯΛड͚Δslaveαʔό্ͰͷTriggerɺStatementํࣜͷ߹ͷ Έ࣮ߦ͞ΕΔ • SlaveͰtriggerΛ҆શʹར༻͢ΔͨΊʹҙਤ͠ͳ͍ Row basedͳϨϓϦέʔ γϣϯΛൃݟ͢Δඞཁ͕͋Δ
• check-mysqlbinlog-rbr-fallbacks • No OSS • mysqlbinlog ίϚϯυ͍ɺRow BasedʹͳͬͨςʔϒϧΛݕग़
Master Backup Anon Slave Trigger binlog binlog check-mysqlbinlog-rbr-fallbacks Detect the
SQL and fix data and applications if necessary App
• ઓ͍ͯ͠ΔݶΓɺࢹ૿͑Δ • Application Server: 175 Metrics • DB Server:
245 Metrics
• Ҏ্