Upgrade to Pro — share decks privately, control downloads, hide ads and more …

SRECon19 AsiaPacific Recap

SRECon19 AsiaPacific Recap

Avatar for Takeshi Kondo

Takeshi Kondo

August 02, 2019
Tweet

More Decks by Takeshi Kondo

Other Decks in Technology

Transcript

  1. What is SRECon? https://www.usenix.org/srecon gathering of engineers who care deeply

    about site reliability, systems engineering, and working with complex distributed systems at scale
  2. What kind of talks? • Main • Postmortem, Retrospective •

    SLO • Availability • Organization, Technical Leader, Onboarding • Monitoring, Observability • MLOps • Cloud Billing • Release Engineering • Design Doc • Stress Test, Capacity Planning • Microservices • Scale Database(Kafka, Cassandra) • gRPC • Core • Distributed Systems • Distributed File System • Security Control • Elasticrsearch • Python Global Interpreter Lock(GIL) • Distributed Consensus • Edge Computing • Networking, TCP, BGP • Memory Management • ARM64 • Java Garbage Collector • HBase
  3. Let’s Join • SRECon19 EMEA 2–4 October, 2019, in Dublin,

    Ireland. • Proposals for talks due: May 21, 2019 • Lightning talks: Thursday, August 22, 2019 • SRECon20 Americas 24-26 March, 2020, Santa Clara • SRECon20 AP 15-17 June, 2020, in Sydney, Australia
  4. Recap • Leading without Managing: Becoming an SRE Technical Leader

    / Todd Palino, LinkedIn • Cross Continent Infrastructure Scaling at Instagram / Sherry Xiao, Facebook
  5. Leading without Managing: Becoming an SRE Technical Leader • Summery

    • Leadership is important but difficult to measure • Because human does not scale, leadership is an effective means to maximize the effect • Takeaways • Mentoring • Looking to the Outsize • Keep A Paper Trail • Impression • Important to influence and lead technically • Both for the organization, for the individual, both are worthwhile • Ability required for SRE
  6. Leading without Managing: Becoming an SRE Technical Leader • Summery

    • Trade-offs for the reliability of data across continents • Takeaways • If you really care about latency and performance, Bring your data closer to a user • To make your service disaster resilience and be able to feel over, you should deploy to multiple regions • And you should partition data set to reduce replication, this will help you save a lot of storage space. • You can't have all you have to make trade offs • Impression • Large-scale cases are interesting
  7. See also (Thanks @chokkoyamada ) • SREcon19 Asia/PacificࢀՃϝϞ: 1೔໨ •

    https://road288.hatenablog.com/entry/2019/06/13/085823 • SREcon19 Asia/PacificࢀՃϝϞ: 2೔໨ • https://road288.hatenablog.com/entry/2019/06/14/092009 • SREcon19 Asia/PacificࢀՃϝϞ: 3೔໨ • https://road288.hatenablog.com/entry/2019/06/16/011556
  8. Summary • SRECon is a very interesting conference for us

    SRE. Let's participate • Almost all the SRECon talks are public. Does anyone learn from those talks together? • Let’s share our knowledge to world through the SRE Lounge
  9. Thank You! chaspy chaspy_ / chaspy_en Site Reliability Engineer at

    Quipper Takeshi Kondo SRE Lounge Terraform-jp
  10. “Okimochi” from the standpoint of the speaker • ւ֎ΧϯϑΝϨϯεͰ࿩͢ͱ͍͏͜ͱ •

    πφϫλϦϚΠϥΠϑ • https://blog.chaspy.me/entry/2019/06/12/120000 • ӳޠͰͷϓϨθϯΛ΍Γ͖Δͨͬͨ1ͭͷίπ • Quipper Product Team Blog • https://quipper.hatenablog.com/entry/2019/06/21/080000 • I write an English article • Answer the questions • Introduce the SRE Lounge • Thank those speakers
  11. Thank You! chaspy chaspy_ / chaspy_en Site Reliability Engineer at

    Quipper Takeshi Kondo SRE Lounge Terraform-jp