Transition & operate

StageTahap
2 · Lifecycle
Reading timeWaktu baca
9 min9 menit
LevelLevel
BeginnerPemula

"Go-live" is a single moment. Transition is all the work that makes that moment safe, and operate is everything that keeps the service running afterwards, day and night.

"Go-live" hanyalah satu momen. Transition adalah semua pekerjaan yang membuat momen itu aman, dan operate adalah segala hal yang menjaga service tetap berjalan setelahnya, siang dan malam.

Transition covers assessing risk, release, deployment, training and handover up to go-live and early support. Operate starts at go-live and covers monitoring, maintenance, patching and backups.TransitionAssessriskReleaseDeployTrainHandovergo-liveOperateMonitorMaintainPatchBack upearly life support overlaps both
Transition leads up to go-live. Operate starts there and never really ends.Transition mengarah ke go-live. Operate dimulai dari sana dan tidak pernah benar-benar berakhir.

The two activitiesDua aktivitas

Transition

Move new or changed products and services into live use, while protecting everything that already works.

Memindahkan product dan service baru atau yang diubah ke penggunaan nyata, sambil melindungi semua yang sudah berjalan.

Typical workPekerjaan umum
Assessing risk and authorizing the change, scheduling, release and deployment (often in phases), communication and training, handover to operations and support, early life support.Menilai risiko dan mengotorisasi change, penjadwalan, release dan deployment (sering bertahap), komunikasi dan pelatihan, serah terima ke tim operasional dan support, early life support.
OutputsHasil
A live change, updated records, a trained support team, and a tested rollback plan.Perubahan yang sudah live, catatan yang diperbarui, tim support yang terlatih, dan rencana rollback yang sudah diuji.
AskTanyakan
What could break, who needs to know, and how do we undo it?Apa yang bisa rusak, siapa yang perlu tahu, dan bagaimana membatalkannya?
Practices that helpPractice yang membantuChange enablementRelease managementDeployment managementOrganizational change management

Operate

Keep live products, platforms and their components running as agreed, around the clock.

Menjaga product, platform, dan komponennya yang sudah live tetap berjalan sesuai kesepakatan, sepanjang waktu.

Typical workPekerjaan umum
Monitoring and handling events, routine maintenance and patching, backups, managing capacity, security operations, automating routine tasks.Memantau dan menangani event, pemeliharaan rutin dan patching, backup, mengelola kapasitas, operasi keamanan, mengotomasi tugas rutin.
OutputsHasil
A stable service, operational data and early warnings.Service yang stabil, data operasional, dan peringatan dini.
AskTanyakan
Would we notice a problem before users do?Apakah kita akan menyadari masalah sebelum pengguna menyadarinya?
Practices that helpPractice yang membantuMonitoring and event managementInfrastructure and platform managementCapacity and performance managementAvailability management

Smaller steps, safer changesLangkah lebih kecil, perubahan lebih aman

The safest transitions are small and reversible. Release to a few users first, watch what happens, then widen. Decide in advance what "healthy" looks like, such as error rate, response time or completed transactions, so the go or no-go call is based on data rather than hope.

Transition yang paling aman berukuran kecil dan bisa dibatalkan. Rilis ke sedikit pengguna dulu, amati hasilnya, lalu perluas. Tentukan sejak awal seperti apa kondisi "sehat", misalnya tingkat error, waktu respons, atau transaksi yang berhasil, agar keputusan lanjut atau berhenti didasarkan pada data, bukan harapan.

Key ideaInti materiA good transition makes operate boring. Boring is exactly what users want from a live service.Transition yang baik membuat operate terasa membosankan. Membosankan adalah hal yang justru diinginkan pengguna dari service yang live.
Real-world example: an e-wallet changes its payment gatewayContoh kasus: dompet digital mengganti payment gateway3 min3 menit

An e-wallet moves to a new payment gateway. Instead of switching everyone at once, the team routes 5% of transactions through the new gateway, then 25%, then 100% over two weeks. A switch lets them send traffic back to the old gateway in seconds.

Sebuah dompet digital pindah ke payment gateway baru. Alih-alih memindahkan semua orang sekaligus, tim mengalihkan 5% transaksi ke gateway baru, lalu 25%, lalu 100% dalam dua minggu. Sebuah sakelar memungkinkan mereka mengembalikan lalu lintas ke gateway lama dalam hitungan detik.

At the 5% stage, the operate dashboards show that payments to one partner bank are timing out. They pause, fix the timeout setting, and continue. Only a handful of users were affected instead of millions.

Pada tahap 5%, dashboard operate menunjukkan pembayaran ke satu bank mitra mengalami timeout. Mereka berhenti sejenak, memperbaiki pengaturan timeout, lalu melanjutkan. Hanya segelintir pengguna yang terdampak, bukan jutaan.

Beginner tipsTips untuk pemula
  • Common mistake: a big-bang release on Friday evening, with no rollback plan and no one on call.
  • Hand over with a short runbook: how to check health, common fixes, and who to call.
  • Pick three health signals before go-live and put them on one screen.
  • Kesalahan umum: rilis besar-besaran di Jumat sore, tanpa rencana rollback dan tanpa petugas siaga.
  • Lakukan serah terima dengan runbook singkat: cara memeriksa kesehatan sistem, perbaikan umum, dan siapa yang harus dihubungi.
  • Pilih tiga sinyal kesehatan sebelum go-live dan tampilkan di satu layar.

Activity names are ITIL’s own. Public overviews don’t yet give full official definitions, so the explanations, examples and practice links here are written in our own words.Nama aktivitas berasal dari ITIL. Ringkasan publik belum memuat definisi resmi yang lengkap, jadi penjelasan, contoh, dan kaitan practice di sini ditulis dengan kata-kata sendiri.