Segment 1: HiddenMerit Morning Post · Issue 88

Focus on Database Frontiers, Practical Insights for DBAs August 28, 2026 ｜ 5 Selected Global Breaking News

01｜Alibaba Cloud OS Console Launches Two Skills: Early Morning Alarms No Longer Panic, Inspection + Diagnosis Closed Loop Goes Live

Alibaba Cloud OS Console has recently released two OS operation and maintenance Skills – SysOM Inspection Skill and SysOM Diagnosis Skill – which will later be merged into the SysOM Skill Package, forming a complete closed loop of “inspection discovers problems → diagnosis locates root causes → remediation recommendations are provided.”

A New Solution for Early Morning Faults: In a real operation scenario, the SysOM Inspection Skill completed 19 inspection items in 37.4 seconds, identified an anomaly with 91.90% memory usage, and automatically linked to diagnostic capabilities, directly locating a Python process occupying 12.95 GB of memory. The entire process – from alert to root cause location to recommendations – was compressed from 40 minutes to under 5 minutes. Currently, the SysOM Inspection Skill covers over 40 inspection capabilities across system load, memory, disk latency, scheduling latency, socket leaks, and multiple types of memory leak risks. In internal evaluations, the overall anomaly identification accuracy exceeded 80%, with a 0% false positive rate for high‑risk items.

DBA Perspective: The essence of the SysOM Inspection Skill is the transformation of senior operations engineers’ “troubleshooting experience” into reusable, automatable, standardised capabilities. When database alerts no longer depend on “who is more alert tonight” but are instead automatically root‑caused by AI, the frequency of early morning wake‑ups for DBAs will significantly decrease. The SysOM Skill Package plans to merge “inspection + diagnosis” capabilities to form a complete closed loop from discovery to remediation. For DBAs, this means a shift from reactive response to proactive prevention, freeing up more energy for architectural design and capacity planning.

📌 HiddenMerit In‑Depth: Today’s follow‑up article – “SysOM Inspection Skill Deep Dive: When Operational Experience Becomes a Reusable Skill” – will detail the technical principles and practical value of this Skill for DBAs.

📌 HiddenMerit Mid‑Day Preview: Today at noon, we will focus on OpenAnolis SkillHub – when OS‑level operational experience is standardised into Skills, can DBAs call system‑level diagnostic capabilities as easily as installing apps?

02｜OpenAnolis Community Upgraded to “AI‑Native OS Community,” SkillHub Platform Launched

On June 25, at the 2026 Open Atom Open Source Ecosystem Conference, the OpenAnolis Community officially announced its upgrade to an AI‑Native Operating System Community and launched SkillHub, a community‑level Skill ecosystem platform for agentic systems.

Strategic Significance: The OpenAnolis Community has evolved from “hosting AI” to “understanding AI, scheduling AI, and collaborating with AI,” becoming one of the few open‑source communities globally that systematically advances both “System for AI” and “AI for System” tracks. SkillHub is the first platform to沉淀 OS‑level operational experience, security policies, and inference scheduling logic as standardised, shareable system‑level Skills. Founding contributors include Alibaba Cloud, ZTE, Inspur Information, Intel, Hygon Information, UnionTech Software, AMD, Arm, and Biren Technology.

DBA Perspective: The Skill ecosystem at the OS layer is echoing the Agent Skill ecosystem at the database layer. In the future, DBAs will face not just database‑level operations, but the coordination of Skills across three layers: “operating system + database + AI agent.”

03｜CVE-2026-14669 (CVSS 8.8): PostgreSQL to_char() Heap Overflow PoC Publicly Released

Full technical details and PoC for the PostgreSQL heap buffer overflow vulnerability (CVE-2026-14669, CVSS 8.8) have been publicly released. The vulnerability stems from the to_char() function failing to perform length checks when processing overly long POSIX time zone abbreviations. Affects all PostgreSQL versions before 18.5, 17.11, 16.15, 15.19, and 14.24. The PostgreSQL team fixed this in the August 2026 update.

DBA Perspective: The CVSS 8.8 rating combined with a public PoC means the exploitation barrier has been significantly lowered. to_char() is a widely used time‑formatting function – any user with SQL execution privileges can trigger it. DBAs are advised to immediately check versions and prioritise upgrades within the Q3 patch window.

04｜Domestic Database “Big Three” Market Share Exceeds 35%, Domestic Replacement Accelerates at Scale

According to the latest CCID Consulting data, the combined market share of Dameng, Renda Kingbase, and OceanBase in China’s relational database market has exceeded 35%. China Merchants Securities analysis suggests that after the domestic database market share breaches the 30% tipping point, it will accelerate into the large‑scale replacement phase, with core system replacement in finance, telecommunications, government, and other key industries becoming the main battlefield.

DBA Perspective: The establishment of the “Big Three” pattern means the priority for Xinchuang skill development is now clear – Dameng, Kingbase, and OceanBase are the three product lines that DBAs must cover.

05｜PostgreSQL 14 EOL Countdown: Only 76 Days Until End of Support

PostgreSQL 14 will officially stop receiving patch updates on November 12, 2026 – only approximately 76 days remain. PG 18 is the current recommended LTS version, with support extending to May 2031.

DBA Perspective: With less than three months until PG 14 EOL, DBAs must immediately initiate upgrade assessments. It is recommended to use pg_upgrade --check for compatibility pre‑checks, with special attention to extension module dependencies and postgresql.conf parameter differences.

📌 This Issue’s Previews

📌 HiddenMerit In‑Depth (with this Morning Post) : SysOM Inspection Skill Deep Dive – When Operational Experience Becomes a Reusable Skill

📌 HiddenMerit Mid‑Day (Today at noon) : OpenAnolis SkillHub First Look – What Does the OS‑Level “Skill App Store” Mean for DBAs?

📌 HiddenMerit Night Talk (Tonight) : The 4 AM Alert – Is the DBA “Wake‑Up” Moment Becoming History?

---

HiddenMerit Team Production Slogan: 绩优隐于内，金石启新程 | Hidden deep. Merit bold. Forge ahead.

Segment 2: HiddenMerit In‑Depth · Issue 1

Focus on Database Frontiers, In‑Depth Analysis August 28, 2026 ｜ 1 In‑Depth Feature Article

SysOM Inspection Skill Deep Dive: When Operational Experience Becomes a Reusable Skill

1. Why Is the 4 AM Alert Always Such a Pain?

At 4 AM, a server memory alert goes off. You wake up, open your laptop, log in to the server, and start troubleshooting:

• top to check process memory usage
• free -h to confirm total memory
• ps aux --sort=-%mem to find the top memory consumer
• Check logs, review monitoring, make a judgment…

Forty minutes later, you finally find the Python process consuming 12.95 GB of memory.

Every DBA has experienced this scenario. What the SysOM Inspection Skill aims to solve is not the “being woken up” itself, but the underlying pattern of “relying on personal experience, relying on improvisation, relying on who is more alert tonight.”

The SysOM Inspection Skill is one of the few open‑source practices that fully encapsulates “operations expert experience” into a reusable Skill. It is not meant to replace DBAs, but to turn DBAs’ “hard‑earned” experience into organisational capability.

2. Technical Principles of SysOM Inspection Skill: 40+ Inspection Items + Automatic Diagnosis Linkage

The SysOM Inspection Skill currently covers over 40 inspection capabilities across the following dimensions:

Dimension	Coverage
System Load	CPU load, run queue length
Memory Usage	Memory consumption, swap usage, OOM risk
Disk I/O	Read/write latency, IOPS, queue depth
Scheduling Latency	Process scheduling latency, soft interrupts
File Handles	Handle count, file descriptor leaks
Thread Resources	Thread count, thread leaks
Network	Socket leaks, tcp/udp memory
Kernel‑Level Risks	Multiple memory leak types, kernel parameter anomalies

Its core design principle is: not to generate more alerts, but to turn alerts directly into actionable conclusions.

• Automatic diagnosis linkage for high‑risk scenarios: When an inspection finds an anomaly, it doesn’t just send an alert notification – it automatically links with the SysOM Diagnosis Skill, covering system load, memory, network, disk, scheduling, and other OS subsystems, tracing the investigation path all the way to the root cause.

• Decision‑maker‑oriented output: The inspection report does not present just a “list of alerts + technical parameters,” but directly answers: “Who did it, how serious is it, and who should be contacted.” In internal evaluations, the overall anomaly identification accuracy exceeded 80% , with a 0% false positive rate for high‑risk items.

• Root cause location no longer depends on humans: The experience chain of “suspicion → investigation → location” becomes a replicable process of “discovery → root cause → recommendation.”

3. Implications for DBAs: When Operational Knowledge Moves from “Head” to “Skill”

The biggest implication of the SysOM Inspection Skill for DBAs is not that “AI can do work,” but that operational knowledge is moving from “stored in the mind” to “encapsulated in Skills.”

In the past, a DBA’s value was reflected in “I remember how to check”:

• I know what to look at first and what to look at next for this alert
• I know what problem this error log corresponds to
• I know the optimal solution for this scenario

Now, this knowledge is being progressively encapsulated into Skills. The SysOM Inspection Skill’s internal evaluation accuracy of over 80% and 0% false positives for high‑risk items prove that these experiences can be systematically replicated.

This means three things for DBAs:

1. Repetitive troubleshooting work will be largely replaced: The scenario of being woken up at 4 AM to investigate root causes will become less frequent, because the Inspection Skill has already done the investigation for you.

2. “Knowing how to do it” is no longer the core competitive advantage: When most standard processes are covered by Skills, a DBA’s core value will shift to “knowing what to do.”

3. Skill definers become a new role: Someone needs to define what counts as “high‑risk memory usage” and what scenarios require automatic diagnosis linkage – these decision boundaries still require human judgement.

4. When Inspection and Diagnosis Form a Closed Loop

SysOM plans to merge the Inspection Skill and Diagnosis Skill into the SysOM Skill Package, forming a complete closed loop of “inspection discovers problems → diagnosis locates root causes → remediation recommendations are provided. ”

Once this closed loop is in place, the DBA’s work model will fundamentally change:

• From “starting the investigation after being woken up” to “having a report waiting when you wake up”
• From “relying on personal experience” to “relying on organisational Skills”
• From “repeatedly troubleshooting the same issues” to “focusing on new problems”

The 4 AM alert may no longer be a DBA’s nightmare – but just another night when Skills automatically complete their tasks.

HiddenMerit Team Production Slogan: 绩优隐于内，金石启新程 | Hidden deep. Merit bold. Forge ahead.

Segment 3: HiddenMerit Mid‑Day · Issue 5

Focus on Database Tools Ecosystem August 28, 2026 ｜ Deep Dive into 1 Hot Tool

This Issue’s Tool: OpenAnolis SkillHub – The OS‑Level “Skill App Store”

When Agent Skills are flourishing at the AI application layer, OS‑level Skills have also arrived.

1. What is SkillHub?

SkillHub is a community‑level Skill ecosystem platform for agentic systems, officially launched by the OpenAnolis Community in June 2026. Its positioning is: to沉淀 OS‑level operational experience, security policies, and inference scheduling logic into standardised, shareable system‑level capability assets.

In simple terms, it encapsulates the expert experience of those who “know Linux kernels, know system operations, know performance tuning” into reusable Skills that AI Agents can directly call.

SkillHub was initiated by the OpenAnolis Community, with founding contributors including Alibaba Cloud, ZTE, Inspur Information, Intel, Hygon Information, UnionTech Software, AMD, Arm, and Biren Technology.

2. What Problem Does SkillHub Aim to Solve?

When upgrading to an AI‑Native OS Community, the OpenAnolis Community identified a critical gap: the Skill ecosystem at the application layer is flourishing, but Infrastructure‑layer Skills have long been scattered everywhere.

Agent applications driven by large models are rapidly proliferating, but agents’ ability to call underlying system capabilities remains highly fragmented. An agent may be good at calling APIs, generating code, and answering questions, but it is not good at:

• Diagnosing a memory leak in the Linux kernel
• Identifying the source of disk I/O bottlenecks
• Determining whether system parameter configurations are reasonable
• Investigating the root cause of network connection anomalies

SkillHub’s goal is to turn the experience of “system experts” into standardised capabilities that agents can also call.

3. What Does SkillHub Mean for DBAs?

For DBAs, the emergence of SkillHub means three things:

1. Operational experience will shift from “personal assets” to “organisational assets”

The SysOM Inspection Skill has already proven: 40+ inspection items, over 80% accuracy, 0% false positives for high‑risk items – these capabilities do not depend on a “guru’s” personal experience, but have been encapsulated into reusable Skills. In the future, a DBA’s “hard‑earned” troubleshooting experience can also be沉淀 into Skills, not lost when people leave.

2. DBAs will shift from “executors” to “Skill definers”

SkillHub supports ecosystem partners in沉淀 operational best practices as standardised capability assets. This means the DBA role may evolve from “repeatedly executing operations” to “defining what operational logic should become a Skill.”

3. The database + operating system + AI operations closed loop is taking shape

When the SysOM Inspection Skill can automatically diagnose root causes at the OS level, database DBAs will no longer need to switch back and forth between databases and operating systems for investigation. Full‑chain observability from operating system to database is being connected by Skills.

4. How to Explore SkillHub

SkillHub is now live, and Skills from various contributing organisations are being released continuously. Directions worth watching include:

• OS diagnostic Skills: automatically diagnose memory, CPU, disk, and network anomalies
• Performance tuning Skills: automatically recommend system parameter configurations based on business workloads
• Security compliance Skills: automatically check whether system security baselines meet compliance requirements

For DBAs, this may be one of the most important skill ecosystems to watch – when OS‑layer Skills and database‑layer Skills are connected, the boundaries of full‑stack operations will be completely reshaped.

HiddenMerit Team Production Slogan: 绩优隐于内，金石启新程 | Hidden deep. Merit bold. Forge ahead.

Segment 4: HiddenMerit Night Talk · Issue 20

DBA Workplace & Life August 28, 2026 ｜ Let’s Talk About DBA Life

Tonight’s Topic: The 4 AM Alert – Is the DBA “Wake‑Up” Moment Becoming History?

1. A True Early Morning Story

At 4 AM, a memory alert goes off. DBA Xiao Lin is woken up.

“At that moment, there was only one thought in my mind: here we go again,” Xiao Lin recalls. “The first thing I did was check memory usage, see who was consuming it, then check logs, review monitoring, make a judgment… by the time I located that Python process consuming 12.95 GB of memory, 40 minutes had already passed.”

In that 40‑minute investigation, his actions were actually very “standardised”:

1. Log in to the server
2. top to check processes
3. ps aux --sort=-%mem to find the top memory consumer
4. Check application logs to verify the hypothesis
5. Remediate after confirming the root cause

Throughout the entire process, the time that truly relied on “human judgment” was less than 5 minutes; the remaining 35 minutes were spent on “standard actions.”

The SysOM Inspection Skill compressed this process to under 5 minutes. No luck required, no reliance on who is more alert tonight – just encapsulating troubleshooting experience into a Skill.

2. The “Craft” of Being Woken Up Is Becoming a Standardised Skill

The tasks a DBA performs after being woken up at 4 AM can be broadly divided into three categories:

Type	Content	% of Early Morning Time
Information Gathering	Log in, check monitoring, review logs, find data	60%
Root Cause Judgment	Analyse data, reason, form conclusions	25%
Remediation Execution	Restart, adjust parameters, contact development	15%

Information gathering and remediation execution are the easiest parts to standardise. The SysOM Inspection Skill is already automating information gathering and root cause location. Standardisation of remediation execution is only a matter of time.

What truly relies on humans is only the part of root cause judgment that requires business context, risk assessment, and decision‑making.

3. What Can Skills Achieve Today? What’s Still Out of Reach?

From the SysOM Inspection Skill’s internal evaluation data: overall anomaly identification accuracy exceeds 80% , with 0% false positives for high‑risk items. This means: when it tells you “there is a high risk here,” it is almost certainly correct.

But what Skills currently cannot do includes:

• Determining “whether this anomaly is worth handling right now” – requires business context
• Assessing “what side effects adjusting this parameter might bring” – requires holistic understanding
• Handling “fault patterns that have never been seen before” – requires creative thinking

These are precisely the values that DBAs cannot be replaced for.

4. The New Battlefield for DBAs: From “Being Woken Up at 4 AM” to “Defining Skills During the Day”

As tools like the SysOM Inspection Skill become more prevalent, the frequency of DBAs being woken up will significantly decrease. But DBAs will not become idle – on the contrary, new work is emerging:

1. Defining the decision boundaries of Skills: What counts as high‑risk memory usage? What scenarios require automatic diagnosis linkage? Who sets the boundaries between “automation” and “human intervention” for Skills?

2. Reviewing Skill execution results: After a Skill automatically locates the root cause, who confirms “is this root cause judgment accurate”? Who evaluates “is this remediation recommendation appropriate for the current business scenario”?

3. Driving Skill iteration direction: Skills can solve 80% of common problems, but who solves the remaining 20% of “long‑tail problems”? Who identifies the capability gaps in Skills and drives iteration?

The 4 AM alert may no longer be a DBA’s nightmare – but a DBA’s work will not decrease; it will just change form.

5. Tonight’s Interaction

Feel free to share your thoughts in the comments:

1. What was the last issue that woke you up at 4 AM? How long did the investigation take?
2. If you could encapsulate your most frequently used troubleshooting process into a Skill, which one would you start with?
3. Would you be willing to contribute your “hard‑earned” experience to become a shared organisational Skill?

HiddenMerit Team Production Slogan: 绩优隐于内，金石启新程 | Hidden deep. Merit bold. Forge ahead.